Skip to main content

tv   Twitter CEO Jack Dorsey Testifies on Content Monitoring  CSPAN  September 5, 2018 11:35pm-3:38am EDT

11:35 pm
c-span, or the free radio app. announcer: sunday night on "afterwards," former obama administration education secretary arne duncan on his book, "how schools work" interviewed by the former chancellor of the district of public schools. >> i don't know if voters make the connection between what and what happens in schools so how do we draw clearly? a little more >> we need voters to understand, access to pre-k, if we want to pay our teachers rates, reduce dropout make college more affordable, we have to challenge and hold accountable our elected officials we put in office. "afterwards"tch sunday night at 9:00 eastern on tv.an2's book announcer: we take you to a
11:36 pm
commercergy and committee hearing. where twitter c.e.o. jack dorsey afternoon.his he explained how the company postings. the hearing lasts about four hours. to order.w come chair recognizes himself for purposes of an opening statement. afternoon and thank you, mr. dorsey, for being before the energy and commerce committee today. and yourny you co-creators founded 12 years ago mostecome one of the recognizable in the world. using the service has become a verb, instantly recognized around the world. just as people can google a a photo,or gram everyone knows what it means to tweet thoughts or ideas. list of superlatives to
11:37 pm
280ribe twitter exceeds characters. it is one of the most downloaded apps in the world and most websites, one of the most premier sources for breaking news. its power and reach are so great that society changing events like the arab spring have been revolution.witter the service allows anyone with access to the internet the power or her views to the world. it's truly revolutionary in the guttenberg press was revolutionary. free,ps set information allows ideas to propagate and challenge established ways of thinking. twitter's success and growth rate has been extraordinary but it is not without controversy. the service has been banned at various times and in various andtries, such as china iran. hear in the united states, the undery itself has come criticism for impeding the ability of some users to post remove tweets and other content moderation
11:38 pm
practices. for instance, in july, it was reported that some politically were no longer appearing as auto populated options in certain search results. to concerns that the service might be, quote/unquote, shadow banning. users, in an attempt to limit visibility on the site. was hardly the first instance of a social media service taking actions which oreared to minimize de-emphasize certain viewpoints. and in the most recent case, twitter has stated that the actions were not intentional of, rather, the result algorithms designed to maintain site. civil tone on the also -- twitter has also directly addressed the automatedots or accounts not controlled by one
11:39 pm
person. even the removal of the bots from the service raised about how the bots were identified because the number of followers someone has economicr has real valley in our economy. we recognize the complexity of trying to manage your service which posts over half a million per day. i believe you were once fromrarily suspended twitter due to an internal error yourself. we do not want to lose sight of few facts. humans are building the algorithms and making decisions to implement twitter's terms of service and implementing changes to policies. humans can make mistakes and how twitter manages those criticallyes is important in an environment are set up toms decide what we see on our feeds. critical that users are confident you are living up to your promises. rules, itto company believes everyone should have
11:40 pm
sharewer to create and ideas instantly without barriers, a noble mission and one private companies do not have to take on. the fact that you have, has enriched the world and changed society, giving an outlet to voices that might not be heard. reassured you are living up to that mission and hope you can help us better when twitter decides when to suspend a user or ban and what the service you do to make sure decisions are made without bias. hope you can help us role algorithms have in this process and how they're designed to ensure consistent and fair outcomes. plays an important role in sharing news and the role.n throughout we appreciate you appearing before us today to answer questions. i yield back the balance of my
11:41 pm
time. >> thank you, mr. chairman. the past few weeks, president trump and many peddledans have conspiracy theories about twitter and social media to whip up their base and fund raise. i fear republicans are using purposesing for those instead of addressing the serious issues raised by social platforms that affect everyday americans' lives. platforms a valuable and can be used to bring people together and allows one to reach many. iran and ukraine, it was used to give voice to concerns of voiceless individuals. closer to home, twitter and hash woke and me too have fosteredlity conversations changing society. users facey twitter bullying, trolling attacks, tweets designed to belittle and individuals can can be devastating, sometimes driving
11:42 pm
to suicide. has takentwitter steps to protect users, more needs to be done. co-opted twitter and other social media platforms sowpread disinformation and divisions in our society. for example, alex jones used amplify harmful and dangerous lies such as those the the sandy hook elementary shooting. it to deny thed holocaust, and spread false terrorism, natural disasters and more. questioned, jack dorsey said the truth will win out in the end but this is reason to doubt that. to a recent study published by m.i.t., the rumors on twitter travel farther and than the truth. that's dangerous. in countries like russia and taking advantage of this to broadly disseminate propaganda
11:43 pm
and false information, beyond elections, foreign agents actively trying to turn groups of americans against each other. they're encouraging confloicket -- conflict-- flict generating topics that create intense feelings. the president using it to words likeople using clown, spoiled, enemies and he routinely creates false statements designed to foster discord and the president's actions feed distrust within our society. president trump has demonstrated the politics of division are good for fund-raising and rousing his base and sadly republicans are following his lead instead of criticizing the president for behavior that tolerated from a child. as reported in the news, the republicanign and majority leader have used
11:44 pm
supposed anti-conservative bias online to fund raise. this hearing appears to be one more mechanism to raise men and generate outrage and it appears republicans are trying to rally their base by fabricating a problem that simply does not exist. regardless of the republicans' for this hearing, twitter and other social media platforms must do more to regain and maintain public trust. bullying, spread of disinformation and malicious influence continue. twitter policy have been inconsistent and confusing. the company's enforcement seems chase the latest headline as opposed to addressing systematic problems. and other social media platforms must establish clear policies to address the problems discussed today, provide tools to users and then swiftly and policies.orce those those policies should apply equally to the president, administration officials, celebrities and the teenager down the street. it's long past time for twitter other social media companies to stop allowing their platforms
11:45 pm
of discord, spreading false information and of foreign government manipulation. thank you for having the hearing. back.d >> thanks, gentlemen. recognizesow mr. dorsey for purposes of an statement. we appreciate your being here. to go mr. dorsey: thank you. thank you, chairman walden, and the committee for the opportunity to speak on behalf of twitter to the people. i look forward to our conversation about our impartiality, to transparency and to accountability. ok with all of you, i'd like to read you something i personally wrote as i thought about these issues.
11:46 pm
i'm also going to tweet it out right now. start by making something very clear. we don't consider political perspectives or party affiliation in any of our enforcement decisions, period. guidinglity is our principle. let me explain why. we believe many people use a digital public square. they gather from all around the world to see what's happening and have a conversation about what they see. rightly serve as a public square if it's constructed around the personal opinions of its makers. we believe a key driver of a square is thec fundamental human right of anddom of opinion expression. our early and strong defense of exchange has enabled twitter to be the activists,r marginalized communities,
11:47 pm
whistleblowers, journalists, and the most influential people around the world. twitter will always default to free exchange. a default to free expression unchecked can generate risks and dangers for people. twitterortant distinguishes between people's andions and their behaviors disarms behavior sphwonged silence another person or adversely interfere with their universal human rights. we build our policies and rules with the principle of impartiality. criteria, rather than on the basis of bias, prejudice, the benefit of one person over another for improper reasons. if we learn we fail to create impartial outcomes, we to fix.ely work in the spirit of accountability recently weency, failed our intended
11:48 pm
impartiality. were unfairly filtering 600,000 accounts including some members of autoess from our search complete and latest results. we fixed it. happen?it our technology was using a decision making criteria that behavior of people following these accounts. andecided that wasn't fair we corrected it. we'll always improve our technology and algorithms to drive healthier usage and impartiality of those outcomes. algorithms is an important topic. our responsibility is to measure, and reduce due to factors such as the quality of the data used to train our algorithms. an extremely complex challenge facing everyone applying artificial
11:49 pm
intelligence. our teams are experimenting with sure thes to make standard is high upholding algorithmic fairness. data, we analyzed tweets sent by all members of founduse and senate and nestatisticcal difference between the number of times a viewedy democrat is versus republican even after all our ranking and filtering of applied.s been also, there's a distinction we need to make clear. when people follow you, you've earned that audience. and we have a responsibility to see your they can tweets. we do not have a responsibility, amplify yourht, to tweets to an audience that doesn't follow you. algorithms decide to show in shared spaces like onrch results is based thousands of signals that
11:50 pm
constantly learn and evolve over time. of those signals are engagement. some are the number of abuse reports. of these toll prevent gaming our system. we acknowledge the growing concern people have of a power like twitter.ies we believe it's dangerous to ask twitter to regulate opinions or the arbtrar of truth. we'd rather be judged by the impartiality of outcomes and criticized when we fail this principle. in closing, when i think of our my mom and dadf in st. louis, a democrat and republican. andad a lot of frustrating heated debates but looking bab, able to heari was and challenge different perspectives and i also so.eciate i felt safe to do we believe twitter helps people connect to something bigger than themselves, show all the amazing
11:51 pm
things that are happening in the world, and all the things we acknowledge and address. we're constantly learning how to make it freer and healthier for to participate. thank you all. >> thank you, mr. dorsey. we'll now begin the opportunity to have questions and i'll lead off. i want to get to the heart of why we're here today. we have a lot of questions about twitter's business practices including questions about your althrights, content management practices and how section 230 twitter.or protects in many ways for some of us, it like theittle bit wizard of oz and we want to know what's going on behind the curtain. reports surfaced that profiles of twitter users drop-downppearing in search results. this was after a member of this her tweets and ads taken off the service because of basic conservative message and other examples sent our way.
11:52 pm
is, ir's public response quote, we do not shadow ban, you're always able to see tweets followcounts you although you may have to do more work to find them like go their profile, close quote. most people might think of that because itanning doesn't matter your definition of shadow banning, when the toectation you are giving your users who choose to follow certain accounts is different line and inime searches. in one example of many, certain usersent conservative including some of our colleagues who have come to us -- jordan,tative meadows, gates -- were not shown in the automatically populated twitter, searches on correct? out of the more than 300 million twitter users, why did this only happen to certain accounts? in other words, what did the algorithm take into account that to prominent conservatives including members of the u.s.
11:53 pm
house of representatives not auto searched in suggestions. what caused that? mr. dorsey: thank you for the question. we use signals, usually hundreds of signals, to determine and to what to show what, to down-rank or potentially what to filter. as iis particular case, mentioned in my opening, we were the behaviorl of of the people following accounts didn't believe, upon further consideration, and also seeing the impact, which was about 600,000 accounts, pretty broad base, that that was ultimately fair and we decided to correct it. we also decided that it was not to use a signal for filtering in general and we that withinorrect search, as well. and it is important for us to, able to experiment toely with these signals and
11:54 pm
have the freedom to be able to them and also remove them because that's the only way we're going to learn. make mistakes along the way and the way we want to be judged is making sure that we recognize those and that we correct them and what we're whetherfor in terms of we made a mistake or not is this of impartiality and specifically impartial outcomes realized that in this particular case and within search, that we weren't driving have done acould better job there. could bots game the system or work to block or silence certain voices, otherwise?r mr. dorsey: we are always looking for patterns of behavior amplify information artificially and that includeion could actions like blocking so that's why it's important that we don't use one signal but we use hundreds of signals and balance
11:55 pm
them accordingly. a perception that a simple report of a term of a servicen of a terms of will result in action or downranking. that is not true. that we use and weigh according to together signals. mr. warden: i have one final question. followers i have and one asked why twitter relies to reporty on users violations. mr. dorsey: this is a matter of scale so today in order to tweets or to remove reports, we do require a of the violation and that report is reviewed by an individual. reports are prioritized based on the severity of the aport so death threats have higher prioritization than all others and we take action on them much faster. algorithms that are
11:56 pm
searching proactively the network and behaviors on the network and filtering and accordingly. that means it might filter behind an interstitial which is graphic or element within our app or service that one can tap or show moreweets replies so in some cases we are proactively, based on algorithms, hiding some content, causing more friction to see it and those are models we learn evolve, as well. >> now recognize the gentleman from new jersey. thank you, mr. chairman. twitter's effect on american raises issues but that's not why the american public has called you here today, mr. dorsey. height of's the hypocrisy that president trump and other republicans criticize twitter for supposed bias when the president uses the platform for juvenile tweets
11:57 pm
and spreading lies and misinformation to the country and the world. opinion, you have an obligation to ensure your platform does no harm to our country and democracy and the american public and one persistent critique of twitter by civil rights advocates and victims of abuse and others is policies are unevenly enforced. rich and powerful get special and others get little recourse when twitter fails to protect them unless the company gets bad press. you've admitted twitter needs to job explaining how decisions are made especially by content moderators. how many human content employors does twitter in the u.s. and how much are they paid? mr. dorsey: so we want to think this problem not in terms of the number of people but how make decisions to invest in versusg new technologies hiring folks.
11:58 pm
>> let me issue these three and ifns on this point you can answer, i appreciate it, if not, if you could get back to us. how many human content moderators does twitter employ in the u.s. and how much are and how much hours of training is given to them to consistency in decisions specifichey given instructions to ensure arebrities and politicians treated the same as everyone else. mr. dorsey: we'll follow up with you on specific numbers but on last point, this is a very important distinction. believe we need to do more around protecting private public figures. i don't know yet exactly how manifest but i do believe it's important that we extend the protection of our privatere to individuals necessarily than public figures. >> i appreciate that because i everyone should be treated the same and you seem to be
11:59 pm
saying that but we have to make enforcement mechanism is there so that's true. reportask if you could back to the committee within one month of what steps twitter is taking to improve the of its enforcement and the metrics that demonstrate improvement, if you could, within a month. is that ok? all right. let me turn to another issue. only have a minute. other technology companies like facebook will commit audits in response andome on this committee these audits seek to uncover how platforms and policies have been to stoke racial and religious violence. ofen the haphazard approach twitter towards developing and enforcing policies, i think your should take similar action so let me ask you these three questions and if you get back tor them, us within a month.
12:00 am
will you commit to working with an independent third party civilution to conduct a rights audit of twitter, yes or mr. dorsey: we will, and we do do that on a regular basis. >> i am asking for an independent third-party institution to conduct a. mr. dorsey: let us follow up. making the commit to results available to the public, including all recommendations and findings? mr. dorsey: yes, we believe we need more transparency around our decisions. , willh your permission you commit to change twitter's policy programs and processes to address these areas of concern? mr. dorsey: we are always looking to evolve our policies. yes. >> mr. chairman, through you if we could get a report back within one month of the steps that mr. dorsey is taking, i
12:01 am
would appreciate it. >> now turned to former chairman of the committee for questions. you, mr. chairman. mr. dorsey, i think it is fair to say that even looking at my twitter feed there are fairly ugly things on twitter that come every now and then. and my is fred upton initials are probably use then -- used more than any other. [laughter] >> might even think that it is bipartisan on old -- on both sides of the aisle. i would like to see civility brought back into the public discourse. in a july post, twitter acknowledged that tweets from bad faith actors who intend to manipulate or divide the conversation should be ranked lower. the question is, how do determine whether a user is tweeting to manipulate or divide conversation? mr. dorsey: this is a great question and one that we have
12:02 am
struggled with in the past. we recently determined that we needed something much more tangible and cohesive in order to think about this work. we have come across health is a concept. we have all had experiences where we felt we have been in a conversation that felt a little bit more toxic and we wanted to walk away. we have all been in conversations i have felt really empowering and something we are learning from and we want to stay in them. right now we are trying to determine what the indicators of conversational health are and we are starting with four indicators. one is what is the amount of conversation -- what percentage is focused on the same things? what is the percentage of shared fax that the conversation -- facts that the conversation is hitting? what percentage of the conversation is receptive? is there a variety of
12:03 am
perspective within the conversation, or is it an echo chamber of the same sort of ideas? we are currently trying to figure out what those indicators measureh are and to them. we intend not only to share what those indicators are that we found, but also to measure ourselves against it and make it public so we can show progress. we do not believe we can really fix anything unless we can measure it. we are working with external parties to help us do that because we know we cannot do this alone. . you believe twitter's rules are clear on what is and is not allowed on the platform? mr. dorsey: if you go to our rules today and sit down with a cup of coffee you would not understand them. i believe we need to do a better job not only with those rules but our terms of service. we need to make them more approachable. we would love to lead in this area and we are working on this. i think there is a lot of confusion around our rules and
12:04 am
also enforcement and we intend to fixed it -- fix it. users friend or someone that they follow grandpa mission to access to that user's personal information to a third party? mr. dorsey: no. if you are sharing your password of your account with another, then they have the rights that you would have to take on that account. the chair now recognizes gentleman from new york. he is not here. we are going by the order we were given. , mr. chairman. mr. dorsey, thank you for
12:05 am
joining us here today. these are important issues. even though the democrats have that some of the reasons why you came our, we think our political, nonetheless, there are some real issues with twitter that i think we can discuss today. as you said, twitter really has become a total for engagement across society and recently we saw some of its positive social change with the role it has movement.the me too nonetheless, twitter has also experienced its own sexual harassment problems to confront. i just wanted to ask you some questions about how twitter is dealing with these issues. i don't know if you are aware of the amnesty international report called toxic twitter, a toxic place for women. are you aware of that? mr. dorsey: i am. >> i would ask your consent to put this in?
12:06 am
it describes the issues women face on twitter and how twitter could change to be more friendly to women. i assume you have talked to amnesty international about this report and some of the rat -- some of the recommendations? have noty: i personally, but i imagine the folks on our team have. we can follow up. detail of thehic types of abuses that have been experienced on twitter, including threats of rape, bodily harm and death. some have found to violate twitter's guidelines, but others were not. probably you and your staff agree that twitter needs to do a better job of addressing instances were some of the users are using the platform to harass and threaten others. so, i am wondering if you can tell me, does twitter currently ofe data on reports of abuse conduct, including on the basis
12:07 am
of race, religion, gender or orientation, harassment or threats of violence? secondly, does twitter have data on the actions it has taken to address these complaints? mr. dorsey: first and foremost, we do not believe that we can create a digital public square for people. safe to --not fill feel safe to participate. this is our number one in singular objective is a company is to increase the help of this public space. we do have data on all violations we have seen across the platform. the context of those violations as well. we do intend, and this will be an initiative this year, to create a transparency report that will make that data more public so that all can learn from it and we can be held publicly accountable. >> that is good news. you say you are how that this year? mr. dorsey: we have a lot of
12:08 am
work to do to aggregate all of the data. also taking action to address some of the decisions that have been identified in this report? mr. dorsey: we are. -- one otherng point i wanted to make is we do not feel it is fair that the victims of abuse and harassment have to do the work to report it. our system does work on reports, especially when it has to take content down. abuse reports is something we look at. it is something we want to go down, not only because we think we can reduce the amount of abuse, but we can actually create technology to recognize it he for people have to do the reporting themselves. series of enforcement action all the way to the extreme of it, which is removing content.
12:09 am
>> mr. chairman, i want to save for the record, i don't think these issues are unique to twitter. i would like for many of the invented borderline conspiracies, i believe this is a real threat and i appreciate you mr. dorsey taking this seriously and your entire organization so that we can really reduce these threats online. i yield back. >> chair recognizes gentleman from illinois. >> mr. dorsey, i am from the same metropolitan area. colin behind you will has been known to be in this committee remake couple times. we are glad to have him back. i want to go to my questions and then hopefully have time for some summation. while listening to users is important, how can anyone be sure that standards about what distracts or distorts are being handled fairly and consistently?
12:10 am
is, doesn't those give power to the loudest mop and ultimately failed to discuss controversial speech? mr. dorsey: this goes back to that framework that we were discussing. are theknow if those right indicators yet, that is why we are looking for outside help to make sure we are doing the right word. we should have an understanding , a tangibleement measurement of our affects on our system. specifically in these cases, we are looking for behaviors that tried artificially amplify have our systems in some way. >> a bought, you would consider that as manipulating the system? mr. dorsey: if a bot is used for manipulation. users bandut if together, is at manipulation? mr. dorsey: the same, that is why it makes this issue complicated because sometimes we see bots, and sometimes we see
12:11 am
human coordination. twitter has a verification program where abuse can be verified by twitter as legitimate. verified users have a blue check mark next to their name on their page. review process for designating verified users aligned with your community guidelines or standards? very frank,to be our verification program right now is not where we would like it to be. we do believe it is in serious need of a reboot and a reworking. it has a long history, it started as a way to verify that the cdc account was the actual cdc account during the swine flu. we broadened it without as many strong principles as we needed, then we open the door to everyone.
12:12 am
unfortunately that has caused some issues. the verified batch also is a signal that is used to rank inject shared areas. >> that was my next question. we do have signals that do that. we are identifying those and asking ourselves whether that is still true and still correct today. >> i am just going to end with my final minute to talk about industry standards. i think -- i think my colleague a crossecause this is that technological you are not the only one trying to address these type of concerns. many industries have banded together to have industry standards by which they can comply and also help self correct.
12:13 am
i would encourage the tech sector to start looking at that motto. there is a lot of them out there. i was fortunate to get this book in one of my visits to tech world. mention fairness, reliability, privacy, inclusion, transparency and accountability as standards that should go across the platform. we need to get their for the use of the platforms and trust. i yield back. >> chair recognizes gentleman from texas for questioning. >> thank you, mr. chairman. for being, thank you here. i am pleased twitter is taking steps to improve the experience on its platform. twitter's policy still leaves consumers in danger of misinformation and harassment. users areto ensure
12:14 am
protected from fake accounts, misinformation and harassment. i know that is an issue you are trying to address. i would like to start off by addressing privacy. twitter changes policy in regards to the general data protection regulation that went into effect by the european union. the gdp are makes it clear that consumers need to be in control of their own data and understand how their data is being given to others. thedorsey, as it stands, united states does not mandate the settings or in force. however, i think they are important for an integral part of consumers. my question is will twitter commit to allowing users in the united states have the options of opting out of tracking despite the fact that there is no current regulation mandating this for protection for consumers? before it wasen enacted and re-complied with
12:15 am
regulation, a year prior we were actively making sure that the people that we serve have the controls necessary to opt out tracking across the web, to understand all of the data that we have inferred on their usage, and to individually turn that off and on. we took major steps to make sure we complied with gdp are -- gd r pr. we are different from our peers. what is on twitter is public. approach andferent different needs, but we do a fundamentaly is human right. we will work to protect it and continue to look for ways to give people more control and more transparency around what we have. >> one of the steps twitter has taken to protect consumers and has been to come together with other social media platforms to
12:16 am
create a global internet form to counterterrorism. there is no forum to counter accounts on social media platforms. what steps is twitter making to work together with social media platforms to combat fake bot that were linked to russian and iranian disinformation campaigns? mr. dorsey: this is a definite -- this is definitely a complicated issue. things.e a few we would love to generally be bots across the platform. we can do that by recognizing when people come in our api. aere are other attacks were people script our website or apt to make it look like they were human so they are not coming through our api. answer, butsimple we have got a better in terms of identifying and challenging
12:17 am
accounts. we identified 8 million to 10 million accounts every week and challenge them to determine if they are human or not. we have ported over half a million accounts every single day from even logging into twitter because of what we'd -- whatt be suspicious we detected to be suspicious activity. we do have a good start. we always want to side with more automated technology that recognizes behavior and pattern, instead of going down to the surface area of names or profile images or whatnot. we are looking for behaviors and the intention of the action, which is often times to artificially amplify information and manipulate others. >> i know it -- i know i am out of my time. chair and recognizes gentleman from texas. -- chernow recognizes gentleman from texas for four minutes. you.ank
12:18 am
i will just say that twitter is, in addition to everything else, a news source. that is how i learned of the death of osama bin laden when seal team six provided that information and a happen in real time late sunday night. twitter provided the information. this morning, sitting in anference, not able to get to television, one of my local television stations was attacked and twitter provided the real-time information and updates, which is extremely useful and that is a tool. i think you. thank you. may give mccain's husband complained a lot over the weekend because every doctored image of meghan mccain that was put up on twitter and it seem like it took forever for that to come down. is there not some way that people can -- i understand the outer rhythms, i understand you
12:19 am
have to have checks and balances, but it should not take hours for something that is that egregious to be addressed. absolutely. that was unacceptable and we do not want to use our scale as a use here. to do two things. number one, we cannot place the burden on the victims. we need to build technology so we are not waiting for reports. we're actively looking for instances. while we do have reports of all changes inng building at technology, we need to do a better job at prioritizing, especially any sort of violent or threatening information. case, thisticular was an image that we did not apply the image filter to recognize what was going on in real time. we did take way too many hours and we are using that as
12:20 am
a lesson in order to help improve our systems. i'm sure you have, but have you apologized to the mccain family? mr. dorsey: i have not personally, but i will. >> among the same lines, but may be different, several members of congress have been affected by what was described as -- does -- is itave to report only fixed if someone complains about it? and no one complained it would have been fixed? mr. jordan, mr. meadows, and their accounts being diminished, is it only because they complained? mr. dorsey: it is a completely fair point. we are regularly looking at the outcome of our algorithms.
12:21 am
it was not just the voices of ambers of congress, we saw general conversation. sometimes we need to roll these and see what happens because we will not be able to test everything while come in the right way. lot of feedback and conversation and that is what prompted more digging and an understanding of what we were doing and whether it was the right approach. as a committee can we expect follow-up to your own internal investigations? is that something you can share with us as you get more? mr. dorsey: we would love to. we want to put a premium on transparency and how we can give you information that is clearly accountable to changes. that is why we are putting the
12:22 am
majority of our focus on this particular topic into our transparency report. we would love to release it. it will require a bunch of work. we would love to. >> i will yield back. >> chair recognizes gentleman from pennsylvania. >> mr. dorsey, welcome. read a few quotes about twitter's practices and i want you to tell me if they are true or not. "social media is being rigged to censor conservatives> ." mr. dorsey: no. "it should looks to me like they are censoring people and they ought to stop it." "twitter shadow banning prominent republicans." is that true? mr. dorsey: nope. these were's -- no. by devinwere made
12:23 am
nuñez and president trump on twitter. i want to place those statements into the record. i think it is important for people to understand the premise of this whole hearing, and the allnt twitter, somehow with the other social media platforms got the singular honor to sit in front of this committee is because there is some implication that your site is trying to censor conservative voices on your platform. now, when he tried to explain as ihadow banning, understand it, you had a system where if people who were following people had some behaviors, that was the twitter -- trigger that allowed you to do the shadow banning. you are really like any core opportunity shadow banner. you did not just shadow band for conservative republicans, you shadow band 600,000 people across your entire platform,
12:24 am
across the globe who had people following them that had certain behaviors that cause due to downgrade them coming up. is that correct? mr. dorsey: correct. was never target at conservative republicans, it was targeted to a group of 600,000 people because of the people who followed them. you determine that was not fair and corrected that practice for it is that correct? mr. dorsey: correct. >> just for the record, since you have been singled out as the social media platform for this committee, twitter under no selectively censored conservative republicans are conservative voices on your platform. mr. dorsey: correct. >> let the record correct that because that is the whole reason we are supposedly here. because housley's and deciliter kevin mccarthy said he thinks this is going on a we think your committee should investigate it and it is a load of krapp. -- crap.
12:25 am
what are you doing to address the will concerns many of us have about twitter. people that use twitter to bully or threat and people, we know that this has led to many prominent users, particularly women who have been targeted with sexual threats leaving twitter because of the toxic environment. i understand you are working with these directions and want to include machine learning but i am concerned these solutions will take too long to deploy and they cannot cure the ills that twitter is currently suffering from. we be assured that you and your company have the proper incentives to address the toxicity and abusive behavior on your platform given twitter's current state? foremost,: first and our singular objective as a company right now is to increase the health of public conversation. we realize that will come at short-term costs.
12:26 am
we realize that we will be removing accounts. we realize that it does not necessarily go into a formula where i think there is a perception that we are not going to act because we want as much activity as possible. is like an economic disincentive to act because it takes people from your platform. mr. dorsey: that is not true. we see increasing health of public conversation as a growth sector. take the hard to path and the decisions in order to do so. lot of theseed a during our last earnings call. the reaction by wall street was not as positive. we believe it was important for us to continue to increase the health of this public square, otherwise no one will use it in the first place. >> i yield back. gentlemenecognizes
12:27 am
from texas for four minutes. for want to thank you, sir, being here voluntarily without subpoena or standing and sitting there all by yourself. that is refreshing. i don't know what a twitter ceo should look like, but you do not look like a ceo of twitter with that beard. to reverse the questions that my good friend mr. doyle just asked. so that we kind of get both sides of the question. post your company twitter indicated some democrat politicians were not properly showing up within search auto suggestions. in other words, your company said that your algorithms were somewhat discriminatory against democrats. which democraty
12:28 am
representatives and accounts that were not properly showing up? mr. dorsey: we typically do not identify those as a matter of protecting their privacy. they have not communicated that, but we can follow up with you. >> can you identify how many? mr. dorsey: we will follow up with your staff on that. >> can you personally vouch that that statement is a true that there are democrat politicians who, when he did the auto shirts they did not show up? mr. dorsey: yes, he was over 600,000 accounts. no, no, there are 600,000 accounts affected, how many democrat versus republican accounts? the allegation that we made, the republicans is that your discriminatory against us, against the republicans, your post says that there were some democrat politicians too.
12:29 am
of 600,000, if there were 1000 republicans and 10 still seemst somewhat biased. if it is 50-50 than that is a different ball game. mr. dorsey: we agree that the result is not impartial. that is why we corrected and fixed it. >> c do agree that there were more republicans than democrats. mr. dorsey: i did not say that. it both ways,have sir. it is either 50-50 or one-sided to -- is disproportionately affected. the allegation is that more republicans were affected. mr. dorsey: we do not always have the best the methods to determine who is a republican and who is a democrat. usually it is known because we run as republican or democrat's that is not hard to identify. when it is self identified it is easier, but we are happy to follow up with you. . my chairman keeps whispering
12:30 am
in my ear, i am glad to have a stafford was a chairman of the committee. more oniscriminate anti-conservative versus pro-liberal? no.dorsey: our algorithms do not take into consideration philosophy or viewpoints. understand.ard to we would not be having this discussion if there was not a that yourreement company has discriminated against conservatives, most of whom happen to be republican. mr. dorsey: i believe that we have found impartial outcomes and we intend to fix those and measure them. my time is about to expire and you said he would provide my staff answers with more specificity and i hope you mean
12:31 am
that. thank you for voluntarily appearing. chair recognizes gentle lady from california for four minutes for questions. you for being here, i know it is becoming a long day for you. i want to talk to about anonymous nation. it is best concerned with identifying individual that with to activity of usuals predict consumer behavior. i wonder if that is becoming a distinction without a difference. even when it is not associated with that user's name. toormation is gathered associate with messages or to be. twitter offers geospatial tweets forat has location and names of location -- locationat names including nicknames. researchers have demonstrated this ability.
12:32 am
even though advertising shall not be considered with identifying individual, how is twitter working to ensure that data is not being used by others to do so? first and foremost, the data on twitter is very different than our. companies given that the majority of our data is public by default. do in for a information around people's or their behaviors on the network, we we havehem to see what collected. second, we turn it off. in terms of our data business, our data business is actually focused on packaging up and making real-time public data. we send everyone who wants to consume a real-time stream of public data you know your public
12:33 am
customer process to make sure the intent is still good and proper and also consistent with how they signed up. >> i previously announced in direct thetee to department of commerce to convene a working group of stakeholders to develop a consensus and space definition of block chain. distributed ledger technologies such as block chain have applications in the communication space, ranging from identity verification to sharing. there currently is no common definition that could hinder in its deployment. your previously expressed interest in the broad application of the technology, including potentially any effort findrify identity to misinformation and scams, what potential information do you see for block chain? mr. dorsey: we need to start
12:34 am
with problems that we are trying to solve for our customers and then look at all available technology in order to understand for it to help us accelerate or make the outcomes better. block chain is one that i think has a lot of untapped potential. around distributed trust and distributed enforcement. deep andot gone as understanding how we might apply this technology to the problems we are facing at twitter, but we do have the company thinking about it today. generates revenue through user provider data. your terms of service, you maintain that what's yours is yours and you own your content. i appreciate that but i want to understand more. do users have say if and how and when it is used? is a twitter has the involving set of used with how they can
12:35 am
interact with user con that -- content and how it can be distributed. the hearing demonstrated how the crux of the issue is how content is used a modified to develop assumptions and influence about users to better target ads and individuals. do you believe that users on the data even when the data has modified, used to develop influence or otherwise? mr. dorsey: what was your question? >> do you believe that consumers on their data? mr. dorsey: yes. >> evil when the data is modified or supplemented by additional data or otherwise? mr. dorsey: yes. we would want to understand all the ramifications of that, but yes, we believe people own their data and should have ultimate control over it. >> i yield back. chair now recognizes
12:36 am
-- for four minutes. your. dorsey, appreciate coming and has others have said, we are welcoming your testimony and willingness to answer some of these questions. i think there are serious concerns more than anything where it will be used. really there are many examples of things that twitter has done. many people would express that a for itthe real ability to take off start it with platforms like twitter. in 2009 your band in iran. we have seen other countries, china and north korea banned twitter. when twitter was banned it could not have been a good feeling. but what we are concerned about ,s how twitter looks like
12:37 am
selectively, adversely affected conservatives. i would imagine you are familiar with these. when our calling announced her campaign for the sitter -- senate, twitter quickly banned her announcement advertisement because it had a pro-life message. she at the time was the chair of the special select committee that a number of my colleagues, both republican and democrat were looking into the sale of body parts. twitter band her because -- banned her because they said the statement was inflammatory to invoke a negatively reaction. the you familiar with this? mr. dorsey: yes. forhy was she banned stating a fact that congress was investigating the deep concern nationally when the scandal to place -- took place? >> this was a mistake by twitter
12:38 am
? mr. dorsey: this was a mistake by twitter. >> was anybody held accountable? there was a spokesperson who says we deem it inflammatory, twitter deems it inflammatory. at the same time, the organization that was selling the body parts was not and by twitter, but our colleague who just exposed the fact that the sale of body parts was going on was banned by twitter. one of your own spokesperson said that it was inflammatory. was that person held accountable for making those statements? we use these events and opportunities to improve our process. >> and we have talked about that. obviously, i appreciate the fact you have a knology that there have been mistakes made in algorithms. withve talked about that other companies. facebook talked about similar concerns that we had with their
12:39 am
algorithm and how we felt that against a bias conservatives. a liberal website, vice did a study of all members of congress, all 535 and they identified only three that they felt were targeted in the shadow banning. that was representative meadows, jordans and gates. i know law mr. barton was trying to get into this in more detail, --there was 600,000 ultimately, they did a study and found only three members of congress were bias against and all three happen to be conservatives. can you at least see that that is a concern that a lot of us have if there is a bias in algorithm as it was developed? written algorithms before. if somebody wrote an algorithm with a bias against conservatives, i would hope you are china find out who those people are and if they are you -- trying to find out who those
12:40 am
people are and if they are using viewpoints against people. hope that you would want to know if there are people working for twitter that did have that kind of discriminatory viewpoints against conservatives that you would at least hold them accountable so it does not happen again. thatorsey: i assure you the algorithm was not written with the intention. the signal that we are using caught people up in it. it was a signal we determined was not relevant and also not fair in this particular case. there will be times, and this is where we need to experiment, as you know in writing algorithms in the past, you need to test things and see if they work and pull them back correctly if they don't. >> but also you should not injector own personal viewpoint into that. you are saying it is not. mr. dorsey: that is not the intention. answeringciate you these questions. hopefully we can get more answers to these examples and others like this that we would
12:41 am
truly like to have addressed. --chair now recognizes order. we will have order in the hearing room or you will be asked to leave. , if you will please take a seat or we will have to have you -- then you will need to leave. , mr. presidentus before it is too late. jack dorsey is trying to influence the election and sway the election. >> what is she saying? $30 down here, 35 seven and a half 40. 45, 7 and a half, $50 down here. >> officer will you escort this lady out. >> $80. a hundred dollar.
12:42 am
one and a quarter. two and a quarter. 300. 75, 400. four and a quarter. we are selling a cell phone. five and a quarter, five and a half. i yield back. [laughter] [applause] >> i think our auctioneer is going to get tweeted about today. i would remind members of the audience you are here to observe not participate. i would appreciate that. we will turn to the gentleman from new york for four minutes. >> that is a hard act to follow, mr. chairman. maybe i will get mr. long to help me. thank you mr. chairman.
12:43 am
mr. dorsey, welcome. a directry is facing threat to our democratic institution. we need to find ways to stop throwing adversaries like russia from using american technology against us. special counsel robert mueller filed an indictment against the russian internet research agency think the creative fake social media accounts using american stolen identities to sow discord and interfere with our 2016 election. copy of the indictment here. mr. chairman i would like to introduce it to the record. mr. dorsey, twitter recently took down a number of russian and iranian linked accounts ater it was tipped off by cyber security firm. i am glad to see that twitter is taking action to protect does, but do you think we should be concerned that an outside cyber security firm detect did activity before you did? is really: i think it
12:44 am
important that we have outsiders channel to them because they will always approach the data and work in a way that we may not see. going to do our best to capture everything that we can and to be as proactive as we can. we want to leave room for others to bring a different perspective that might look at what is happening on a platform in a different way than we do. confident are you that twitter can identify and remove all of the fake an automated accounts linked to a foreign adversary? mr. dorsey: we are getting more and more confident. i do want to state this is not something that as an endpoint that reaches perfection. we just have to stay 10 steps vectors attacking newer . we are getting more add tile and better at identifying those and that is showing in some of our results, which i talked about earlier in the terms of being
12:45 am
able to identify 8 million to 10 million suspicious accounts every week. also challenging them to see if the human or bots are malicious automation. >> i understand twitter is acquiring suspicious accounts and aspiring to capture their human accounts and not bots. i was surprised to learn that you are not requiring users to do the same thing when they first sign up for twitter. using only email address, could you tell me why that is? mr. dorsey: we actually do send accounts through a variety of authentic -- of authentication. it really depends on the information we have. we have thwarted over a half a million from marketing. i understand that dealing with foreign adversaries can be difficult.
12:46 am
twitter may respond to one practice only to find new tactics being used to sell -- to sow discord. can you commit to us with any level of certainty that the 2018 midterm elections in the united states will not be subject to interference by the foreign adversaries using bots or other fake accounts on your platform? >> we are committed to making our number one priority to protect the 2018 midterms and the conversation around it. let me ask you this, are you aware of foreign adversaries using any different tactics on your platform to interfere in our 2018 midterm elections? mr. dorsey: we have not -- none tot we have not communicated the senate intelligence. and any we find we will communicate to them. >> we now go to the gentleman from ohio. you fororsey, thank being here with us tonight.
12:47 am
i would like to ask my first how you are protecting that user's data. you collect data from third parties about twitter users? mr. dorsey: we do not collect data from third parties about twitter folks. we do have in beds of tweets around the web. when people do go visit those sites, we note that and we integrate it when they do log into twitter but people can turn that off as well. >> how does twitter use that data? mr. dorsey: we use the data to .ersonalize the experience in my in for a particular interest so we can show specific topics or make our advertising better. sold or offered to other advertisers? is that sold through the advertisers? mr. dorsey: no. to where heck up
12:48 am
was warming to act about the verifications of the blue checkmark. someone toasy for obtain a verified twitter handle and what does twitter take without highlighting one viewpoint over another without verification on a platform? extremely: it is challenging because we have paused the verification program because we found so many full that we needed a read -- faults that we needed a recount. we do make exceptions for particular ramps or public figures of interest, but we generally have paused that were read before that path we did -- we did allow a process. we use criteria to determine if verification was necessary. you have said it can be
12:49 am
removed from the activity of the on/off platform. what platform is the basis for some of what the blue verified check mark? mr. dorsey: we look at specifically any violent extremist groups. off platform behavior for extremist groups not only for verification by holding an account in the first place. in a statement, in the last year twitter developed to launch more than 30 policy and product changes to foster information in integrity and protect the people who use our service for malicious automation. can you share with the committee what those 30 plus policy and product changes are or highlight some? mr. dorsey: we can follow-up with all of you on exactly the details. models, for new instance, to detect where people are gaming our systems.
12:50 am
these are algorithm with an intent to artificial amplified. we have new report inflows that enable people to report tweets or accounts. we have changed policies reflective of current circumstances and what we are seeing. we have certainly done a bunch which hasong gdpr, affected our work in general. we will follow up with you. indicate in your written statement that the company conducted an internal analysis of members affected by the search issue. you make that information available to committee if requested. will you commit to us on the committee that you will give all analysis as soon as possible after this hearing? mr. dorsey: yes. we also hope to include this in our long-standing initiative of transparency report around our actions. >> my time has expired. >> chair recognizes gentle lady
12:51 am
from florida. mr. dorsey, do you feel like you are being manipulated yourself as part of a manipulation campaign? leaderu see the majority of the congress is running ads on facebook to fund raise around allegations of anti-conservative bias on social media platforms, then you see the trump campaign use president trump's tweets where he claims anti-conservative bias at google, facebook and twitter, then we saw the out first today of the woman who jumped out with her phone which will probably spread on the web. but now the justice department is so syria, -- serious, we have to investigate. does the selective manipulation campaign? i do believe there is current concern around the power that companies like ours hold.
12:52 am
the reason why is, people do see us as a digital public square. it comes with certain expectations. >> that is a very diplomatic answer. there are very serious questions. trolls created to influencebots our democracy and elections. they are doing get in other countries across the world. the feel like you have a handle on these bots? you said you id eight to 10 million accounts per month, is that right? mr. dorsey: per week. toward over- and half a million accounts. >> can twitter keep up? mr. dorsey: we intend to keep up. >> if they are using automated accounts, don't we reach a point where they have the ability to overwhelm content on twitter and affect your algorithm? others have
12:53 am
described this as an arms race, but i believe it is right much like security. there is no perfect endpoint. when you build a law, someone else will figure out how to break it. therefore you cannot try to design and optimize for the perfect lock. >> can't you identify the bots as they sign up so that folks understand, that is a fake automated account? mr. dorsey: in certain cases we can, it is a great point. there are more sophisticated ways of automation that actually script our site and our app that are much harder to detect because they are intending to look like human behavior rather than the speed through an api. it is a little bit more complicated. it is not a challenge we are not intending to face. you would have some creative
12:54 am
minds, i would think you could put all of those creative minds and expertise to work to do that. i want to ask you a little bit about privacy, twitter and other companies collect on users, often times without their knowledge. twitter is based off advertising. on fast amount of data that you collect, which raises privacy concerns. , the privacy year policy included a promise of do not track, but then he changed your mind. why? why shouldn't it be up to consumers on tracking? mr. dorsey: we do allow consumers within the app to turn off tracking across the web. >> but you are still able to build a profile on each and every user, isn't that correct? mr. dorsey: if they log into the account, yes.
12:55 am
we allow them to turn that off. >> but i understand that even when they go and change an opt out, you are still collecting data and tracking them? mr. dorsey: i do not believe that is a case but i am happy to follow up with that. recognizes the chairman of republican conference. gentle lady for four minutes. >> thank you, mr. dorsey for joining us. i want to start off by saying i think twitter is a valuable tool in modern communication, that is why in 2011 i was spearheading an effort to get our members signed up and using this tool. i think it is a great way to interact with the people we represent. since then it has been amazing to see the growth of twitter and the twitter users all across america and the world. hearingwhy i think this is so timely. there are serious questions americans have regarding tech
12:56 am
platforms and the ones they are using every day and the issues like data privacy, community standards and censorship. today i want to focus on twitter's procedure for taking down offensive and inappropriate content. as you know, there has been an example that was already shared today, i was going to highlight the one with megan mccain with altered image of a gun pointed at her when she was mourning her father's loss. the tweet image said, america, this one is for you. tweetsly this offensive was reported by other users, nearly you, yet it took 16 hours for there to be action to take it down. i just wanted to ask, do you think this is a violation of twitter's content policies and rules against violence and physical harm, i would also like to understand how much of this is driven by the algorithm versus human content managers? mr. dorsey: it definitely is a
12:57 am
violation and we were slow to act. between tweet was actually up for five hours, that is five hours way too long. model works in terms of removing content the stone reports that we receive. we do not believe that that is fair, ultimately. we do not believe we should put the burden of reporting abuse of harassment on the victim. we need to build algorithms to proactively work for when these things are occurring and take action. the number of abuse reports that we get is the number that we would like to see go down. not only because there is less abuse from the platform, but because our algorithm are recognizing these things before someone has to report them. that is our goal and it will take some time. >> can you talk to me about what are your current policies for prioritizing timely takedowns and enforcement?
12:58 am
any sort of violent threat or image is at the top of our list to review and enforce. we do have privatization mechanism for tweets as we get the reports. obviously the system was put -- was too slow and is not as precise. in this particular reason why is because it was captured within an image within itself. concernsk much of the around it has been how long it takes to remove the content when there is a clear violation. the issue only seemed to be resolved after people publicly tweeted about it. content of a larger people than it ever should've had. i do want to hear what steps the company will take to speed up its response time to ensure these kinds of incidences don't continue? mr. dorsey: in the short term we
12:59 am
need to do a better job at pico -- prioritizing the ones we receive. the longer term we need to take the burden away from the victim from having to report it in the first place. ok, well, clearly you hold a large amount of power in the public discourse. allowing speech that incites violence could have devastating consequences. i believee way where is very important that twitter take action and help restore and yourh the people platform. i will yield back. >> chair recognizes gentleman from maryland for four minutes. dorsey, thank you for coming. there are a number of important topics he could be discussing with you today but the republican majorities decided to pursue the trumped up notion that there is a special conservative bias at work and the way twitter operates.
1:00 am
is allrries me is this part of a campaign by the gop and the right-wing who work for the rats. complaining of nonexistent bias andto force an overcorrection, which can result in actual bias in the other direction. we saw this with facebook. conservatives cried bias because facebook was seeking to make information available using reputable news sources instead of conspiracy platforms. facebook got pushed into this correction. it got rid of its human editors. the result was immediately it was overrun with hoaxes that were posing as news. i actually have questions about the subject of the hearing. i'm going to submit those for the record. and ask for written responses. i do not have confidence this hearing was convened for a serious purpose, to be candid.
1:01 am
i think it is a chance to work awayef to push platforms from the serious task of empowering people with good and reliable information. what is frustrating to me about today's inquiry is that my republican colleagues know there are plenty of other kinds of investigations we should be undertaking in this congress. they do not have any interest in pursuing them. there is a list that has been circulating republicans put together of all the investigations they have been blocking, sweeping under the rug because they want to hide the truth from the american people. this spreadsheet is pretty telling. it is circulating in republican circles. what are these things they know could and should be investigated , but they are determined to dismiss or barry or ignore altogether? according to their own cover-up list, republicans do not want
1:02 am
the public to see president trump's tax returns. they do not want the public to know about trump's business dealings with russia. they are determined not to investigate secretary of treasury steven mnuchin's business dealings. they are blocking public inquiry into the email use of white house staff. they're ignoring how taxpayer money has been wasted by corrupt cabinet secretaries for first-class travel, large security details, expenses and misuse perks. they are giving a president a pass on the motives behind his travel ban and family separation policy. they do not want the public to know how poorly the white house responded to hurricane maria and finally they do not want the public to seeing how the administration is failing to protect our elections and guard against hacking attempts. these are all things that deserve attention and inquiry of congress.
1:03 am
republicans are not going to let it happen. let me talk about election security. we are 60 days away from the midterm election. we know there are ongoing efforts to disrupt our democracy. we know these same actors are using this very platform, twitter and others, to sow discord. we know the public is desperate that their representatives will act to protect democracy. we know thanks to this list the republicans know they should be investigating our nations election security and hacking attempts by hostile actors. instead, here we are using our resources to feed deep state conspiracy theories proffered by the president and his allies in congress. it is a shame this committee has been drawn into such a charade. i yield my time. >> the gentleman's time has expired. the chair recognizes the gentleman from --
1:04 am
>> thank you for taking this time to be here. it is a very important topic. we all utilize twitter. you have a very daunting task to try to work through this. we talked a lot today about algorithms. those are only as good of the people who create them, edit them, guide them. algorithms have to be trained, which means as you know, feeding them a lot of data. oversight of machine learning algorithms involve examining the data sets or the search results to look for that bias. if bias is spotted, the algorithm can be adjusted and retrained. i want to understand the oversight that twitter does of its own algorithms. the algorithms that support twitters algorithmic timeline are adjusted, if not daily, almost daily. why is that? what are some reasons why the
1:05 am
algorithms would need to be adjusted daily? algorithms is a rather new field of research within broader artificial intelligence. it is something that is certainly new to us as a company as well. we have teams to our focused on creating roadmaps so we can fully understand best practices for training data sets and also measuring impartiality of algorithms -- of outcomes. where early in that work. we intend to get better much faster. we are very early. we are learning as quickly as possible, as is the industry, on how best to do this work and also measure whether we are doing the right thing. in terms of why we need to change the signals all the time is because when we release some of these models, we release them
1:06 am
in smaller tests. as they go to be broader twitter at scale, we discover unexpected things. those lead to questions, which then causes us to look deeper at the particular signals we are using. as we recognize there are any sort of impartiality within the outcome, we work to fix. it is dependent on people giving us feedback. >> those teams you are talking about, those are individuals, correct? employees of twitter? >> yes. >> how do you take into account what their leanings or their life story -- does that have an input into what they determine is important? >> it does not have an input we use. the way we judge ourselves ultimately is, are the algorithms making objective decisions? our engineers using
1:07 am
engineering rigor? which is free of bias and any action that might be aligned with one particular perspective or not. >> if i can ask this, what are they looking for? what are they looking for when they are deciding whether or not to make a change? >> they are looking for fairness. they are looking for impartiality. >> if i can interrupt. who defines fairness? what is fairness determined? your fairness may be different than my definition of fairness depending on what the issue or the interpretation of it is. >> this goes back to indicators we are trying to search for. are we showing a variety of perspectives? are we creating more accor chambers and filter bubbles? >> as you look at these users and then specifically you were asked about -- you said you
1:08 am
would follow up on a number of democrats or republicans in the house. my question is, it is a limited pool. we are talking about 435 members of the house. do you have that info and just do not want to discuss it? or do you have to find it? >> we do have the info. we will share it. >> can you share it now? >> we will. >> can you share it now? >> i do not have it with me. >> the chair now recognizes the gentleman from california. dorsey foryou mr. the frankness you have been showing. ais hearing is really desperate effort to rally the republican base before the november election and to please president trump. there are some real serious issues we should be examining. for example, targeting.
1:09 am
some social media networks have been accused of facilitating discriminatory advertising, such as housing and unemployment as -- and employment ads. our advertisers able to exclude certain users on twitter, which would be discriminatory? >> for political ads or issue ads? ads, ournpolitical advertisers able to exclude categories of users? >> advertisers are able to build criteria to include and exclude folks. >> that would end up being .iscriminatory. >> perhaps, yes >> apart from reviewing how advertisements are targeted, this twitter review how ads are delivered? and it discriminatory effects occur as a result of its optimization process? >> we do regular audits of how ads are targeted in delivered and we work to make sure we have
1:10 am
fairness. >> you briefly describe the process twitter uses for making changes to algorithms? >> in terms of making changes to ads algorithms we are looking first and foremost that the data test sets we run through to make sure they are performing in the way we expect with those outcomes. we bring those out to production on the live system. also we are doing checks to make sure they are consistent with constraints and boundaries we expect. >> has twitter ever taken down an ad because of discriminatory effects, nonpolitical? >> i will have to follow up on that. >> it is difficult to know if twitters platforms are having discriminatory effect because there is no way for watchdog groups to examine what is happening for potential biases. twitter announced it is making political ads searchable.
1:11 am
how about nonpolitical ads? is there a way for watchdog groups to examine how nonpolitical ads are being targeted? >> are transparency center is comprehensive. >> inc. you. moving on to privacy. twitters privacy policy states we believe you should always know where your -- what data we andect and how we use it, you should have control over both. most americans really do not know what is happening with their data. there is a saying that if you are not paying for a product, you are the product. you agree? >> i do not necessarily agree. i do believe we need to make more clear the exchange, what people are trading to get a free service. i do not think we have done a great job with that. i believe that is important work and we should clarify more.
1:12 am
>> is twitter running institutional campaigns to inform users about how data is being used? >> not at the moment. we should be looking at that and the incentives we are providing people on the platform. i'm going to follow up on prior questions. if users disable the track stillism, does twitter store previously collected data? or does it erase it? when they opt out. x i believe it is erased. we will have to follow-up. >> can you commit to erasing data? >> yes, but make me -- let me make sure we understand the ramifications of that. >> thank you. i yield back. >> we will now take a five-minute recess and reconvene in five minutes.
1:13 am
1:14 am
1:15 am
1:16 am
1:17 am
1:18 am
1:19 am
1:20 am
1:21 am
1:22 am
>> guests will take their seats. if our guests will take their seats and members, we will resume the hearing. i recognize the gentleman from new jersey for four minutes. >> thank you, mr. chairman. mr. dorsey, i have three areas of questioning. , inhe meghan mccain matter your opinion, the photo had been taken down, if those close to the victim had not complained to twitter? >> what it have been taken down if they had not complained? >> correct.
1:23 am
>> we would have taken it down because we would have seemed -- received reports. our system does work based on reports for takedown. violent. horribly we are all opposed to this type of violence on twitter. regardless of when it occurs. certainly we hope you do better in the future. ,ou state in your testimony bias can happen inadvertently due to many factors such as the quality of data used to train models. in addition to ensuring we are not deliberately biasing the algorithms, it is our rest -- our responsibility to reduce accidental bias. machine learning teams at twitter are learning about these techniques and developing a roadmap to ensure our present and future machine learning models uphold a high standard when it comes to algorithmic fairness.
1:24 am
can you give a time frame as to that wouldht expect receive results that are fair to the american people? conservative and perhaps liberals as well. >> i cannot predict a very precise timeframe at the moment. this is something that is a high priority for us in terms of -- as we roll out algorithms, understanding they are fair and that we are driving impartial outcomes. it is hard to predict it -- a particular timeframe. this is not just a twitter issue. this is the entire industry and a field of research with an artificial intelligence. >> i was asked over the weekend whether this will require regulation by the federal government. we are a committee of jurisdiction. i certainly hope not, but i am sure you can understand that we would like this to occur as quickly as possible because of
1:25 am
the great concern of the american people that they're not be bias, intentional or unintentional. >> i do believe you are asking the important questions, especially as we move more of our decisions to artificial intelligence. we need to understand as we use artificial intelligence for more and more of the things we do that number one, there are unbiased outcomes. number two, they can explain why they made the decision in the first place. >> my third area of questioning, prior to 2016, did twitter have any policies in place to address the use of the twitter platform by foreign governments or entities for the purpose of influencing an election in the united states? i am certainly as concerned as any member of this committee regardless of political party about what happened regarding russia in 2016.
1:26 am
prior to 2016 did you have any policies in place? >> i do not have that data right againsterms of policies foreign actors. we did learn a lot within the 2016 election that impacted both our technology and the policies going forward. >> i do not believe this is a partisan matter. it is intolerable that there was any interference. we hope it never occurs again. thank you, mr. chairman. >> the chair recognizes the gentleman from vermont. >> thank you very much. there are really two hearings going on. one is about that man in the white house who has been accusing, as you have been sitting here, the social media giants of interfering in the election. making this claim even as you are testifying. in fact, recently said the media giants were all in favor of hillary clinton. i will give you a chance to ask
1:27 am
whether the company, twitter, had a policy at the company for either candidate in the presidential election. >> we did not. >> absolutely not, i expect. the second is a job we are not doing. earnedhaving mr. dorsey this good opportunity given his experience. social media platforms are being abused. there are efforts that are being made at twitter. we had mr. zuckerberg here some time ago. efforts being made at facebook to deal with false accounts, to deal with hate speech, which you are trying to deal with. to deal with false information, which is not the kind of thing you want on the digital town square. the fundamental question that this committee refuses to ask itself is whether there is a role for publicly elected
1:28 am
officials to make some of these decisions about how you protect people from hate speech, how you protect people from flat-out false information. you mentioned, mr. dorsey, your company is investigating this. that is a good thing. online, -- bottom line, do you believe this should be something decided company by company or should we have rules for the road and the process that is monitored by elected officials and regulatory agencies? harper earlier asked a very good question. what you determine to be fair or i determine, we may disagree. who is going to be the decider? do you believe ultimately it should be a decision on these important questions of privacy, these matters,
1:29 am
you are trying to contend with about the abuse of your -- should be decided on a company by company basis? or should that be a public discussion in a public decision representatives? >> we want to make it a public conversation. we don't want to have the only healthy public square. we want to contribute to all healthy public conversation. independent of what the government believes it should do, we will continue to make this our singular objective because we believe it is right. we will share our work so that others can learn from it. we will learn from others.
1:30 am
>> this is not an area to as responsible-- as you another companies want to be, ultimately there will be a debate between the president and his vision of what is fair and my vision of what is fair. we now have the ftc, they were designed to address problems when we use dial-up telephones. not doneittee has anything to address the jurisdictional issues in public policy questions. i do not believe we should be leaving it to a private companies. appreciate the- efforts that companies are making. >> the chair recognizes the gentleman from texas for four minutes. get you mentioned
1:31 am
-- welcome. mentioned the trust and safety council within twitter. it relies on the trust and safety council for guidance in evaluating and developing its own committee guidelines. square forhat public free exchange of ideas. you have been honest about your personal biases in the biases of people at twitter. how pervasive are the biases on the trust and safety council? councilome context, the is an external organization of organizations that are focused onare particular issues, such as online harassment, bullying,
1:32 am
misinformation. usse are entities that help give feedback on our policies and also the solutions we are coming up with. entities, are there republican, democrat, green party -- do you know there affiliation politically? >> we have some conservative leaning organizations. we don't add to the council based on ideology. it is on the issues. >> i'm sure that this council operates in a vacuum. what other groups help twitter? anything else out there? >> the trust and safety council is advisory. it makes no decisions for us.
1:33 am
most of our decisions are made internally. we definitely taken but from external folks -- input from external folks. we look at the secular trends of what is going on. we do not treat -- take direction from anything external. >> can you list the mumbles of that counsel -- members of the council? >> they are listed on our webpage. >> i will look that up. i want to turn to back home. as you probably heard, a year ago, southeast texas was fighting floods from hurricane harvey. a recent report from my alma mater highlights how platforms like twitter played an important role in natural disaster recovery. it showed the increased use of mobile devices combined with social media platforms have empowered everyday citizens to
1:34 am
report dangerous situations. they can see people in trouble and report that very quickly. how does twitter prioritize rescue service information during disasters? us, it's hits hurricane season. >> we do prioritize community outreach and emergency services on the platform. we have some really good evidence of this specifically with harvey. 27 million tweets regarding hurricane harvey. systems failed and people used twitter to issue sos calls. we saw as many as 10,000 people rescued from this. this is something we prioritize. we want to make sure we are working with local agencies to make sure that we have a lot of strength there. as a fan of the cardinals, i
1:35 am
theyforgive you for -- hacked into my astros account. i yield back. chaired now recognizes the gentleman from new mexico for four minutes. >> thank you for being here today. is it correct that president trump lost followers because your platform decided to eliminate bots and fake accounts? >> yes. >> during the initial purge of box, who lust more followers? president trump or former president obama? >> i'm not sure of those details. that was a broad-based action aggress -- across all of twitter. >> obama lost 2.3 million followers. president trump lost 320,000 followers. >> i would need to confirm that. >> that it -- is what has been
1:36 am
-- that is what has been reported. in awitter in it since -- conspiracy against former president obama? >> i don't think so. >> i want to commend you on your work. with what was done associated with the evaluation following the 2016 election, which look to some of this work. you note that twitter conducted a conference overview a platform activity related to the 2016 election. i assumed that after your preview you felt that twitter had a responsibility to make changes to the way your platform operates to dress future attempts at election manipulation. is that correct? >> yes. this is our number one priority, to protect the integrity of 2018 elections. >> i might ask unanimous consent
1:37 am
to enter three articles into the record. one from january 19. from april 5. and one from august 21. >> without objection. >> the first article says that twitter admits that there were more russian trolls on its site during the 2016 u.s. of presidential election -- u.s. presidential election. is that correct? >> yes. >> was that an outcome of some of the research? >> that was an outcome of the continued work as we dug deeper into the numbers in 2016. that it also correct twitter has suspended more than 1.2 million terrorism related accounts since 2015?
1:38 am
>> correct. >> how did that work him about? -- come about? >> we have been working for years to automatically detect terrorist accounts and automatically shut that down. that has been ongoing work for years. >> i would hope that this committee would commend your work. twitter remove hundreds of accounts linked to a rainy and and russian political meddling. this was reported august 21. is that correct? >> yes. of anyyou aware significant legislation to protect our democracy and elections? >> i am not aware. >> that is because none of it is happening. we haven't done anything in this congress. 87 million facebook users data was improperly shared with cambridge analytica, this committee heard testimony from
1:39 am
mark zuckerberg in april. are you aware of any significant legislation that passed this committee since then? >> no. >> nothing's happened. mr. chairman, we have not done anything for the people that were impacted by equifax. we should use time to make a difference in the lives of the american people and live up to the commitments this committee has made to provide protections for consumers. i yield back. >> chernow recognizes the gentleman from was virginia for four minutes. >> thank you for coming today. year, the fda commissioner reported that there were offers to sell a legal drugs all over social media -- illegal drugs all over social media. the easy availability of these purchases from illegal drug peddlers is rampant and fuels
1:40 am
the opioid crisis. that yourieve platform has contributed to fueling the opioid crisis? >> first and foremost, we have strong's -- strong terms of service that prevent this activity. we are taking enforcement actions when we see it. >> there was a recent study, just published, that analyzed and withter accounts through several thousands of those and found that there were still 2000 illegal drug sales on your account. -- are your it is against her standards. can you tell me how many of
1:41 am
these sites are still up? >> i would have to follow-up with you on the exact -- >> but they shouldn't be up? >> they shouldn't be. hour, within last the last hour, here's an ad for cocaine on twitter. it's still a. it goes on. that site they can buy cocaine, hair when, mess, ecstasy, percocet. --when have ways of being able to filter that out and it still getting on there. i am astounded that that information is still there. we have the next one. this is on cocaine. here's the next one. here you can get, contact us for
1:42 am
any medicine you want. that doesn't say you have to have a prescription. it's on your site. you said you have ways of checking that. within the last hour, it is still up there. we ran the same problem with facebook. zuckerberg came back to me within two hours and it had all come down. they had missed it. they're all rhythm had missed it. their algorithm had missed it. just come from a state that is targeted from this. we don't need to have our social media promoting the use of illegal drugs. i hope i hear from you that you will be taking them down. is that a fair statement? >> i agree with you. this is unacceptable. >> i would hope that you would
1:43 am
move the same resources that have complicated so much of what this hearing has been about today so that you can focus on this to make her that this doesn't happen again. we wouldn't have to reprimand follow the guidelines that you are so proud about. please, take a good look at it. be serious about this next time. >> i yield back. >> the chernow recognizes the member from iowa. memberink the ranking for having this oversight today. you have exhibited a lot of patience. i commend you for that. there have been a lot of great issues brought up.
1:44 am
i think this discussion today has really demonstrated how important twitter is to our national conversation. the good, the bad, the ugly, all of it. i'm glad that we are shining a light on many issues of concern for americans across the country. many of my colleagues has -- have raised legitimate concerns. the spread of misinformation. that can distort and harm our very democracy. these are all important issues. i want to focus on the issue of online harassment and the use of twitter i teenagers. -- by teenager. social media use by the under 18 population continues to increase. people toow young find friendship and community in ways we cannot of imagine growing up, it also may be creating unimaginable crises for many kids.
1:45 am
has beendia in general used for abusive practices like harassment and cyber bullying. totter has been too slow respond when victims report abuse and harassment. these interactions, which adults for youngrtful only, people these can be devastating. they are still developing. they often plays large importance on the reputations with their peers. we have seen too many tragic stories of what can happen when individuals feel moved to harm themselves in response to online harassment. it should be goal -- a goal for all of us to stop that kind of bullying. of the healthiness of conversations on twitter, are you making any specific changes to the experience of your youngest users? >> yes.
1:46 am
we agree with all of your points. this is one of our areas of focus. around harassment in particular. and how it is used and weaponize to as a tool to silence others. the most important thing for us is that we need to be able to measure our progress around it. and understand that -- whether we making any progress whatsoever. >> there is a minimum age of 13? >> correct. >> does twitter put safety checks on the accounts of teenagers? >> we do have various safety checks. we can follow up with your team on that. >> does twitter do anything to look for indications of harmful or dangerous interactions? >> yes. >> it would be good to know that. has twitter conducted any research with outside organizations to determine how it can best combat online harassment or other harmful interactions?
1:47 am
either for children or teenagers are other groups of people. >> we do this through our trust and safety council. we have an organization that represents youth on digital platforms. >> will you commit to publishing a review with outside organizations to help the body way what more twitter can be doing to protect our kids? >> we have not yet but we will certainly work with our partners to consider this. >> i think your three principles , i think we can put those into effect and operationalize those. i really do appreciate your time. thank you very much for your time. i yield back. >> we recognize the gentleman from kentucky for four minutes. >> thank you for being here today. the floor debate so i have been in the capital
1:48 am
building most of the afternoon. glad to be here. i know that i missed some of your answers. i want to go down the path on a couple of things. many of my constituents use twitter and perceive it to be an open market of ideas. we are here today because some questions have been raised about the rules for posting content and whether some people are restricted in practice. ofill come to a question mandatory adjustment. one major issue starts with transparency and how their data is being collected. have spoken about data a few times already this afternoon. to build on those previous questions, what specific data points are collected on twitter users and with whom do you share them? >> we and for interests around inferfurther -- interests.
1:49 am
we can utilize that information to introduce them to new tweets that may be similar or accounts that may be similar. inference of that data is interest. this is all viewable within the settings of the app. you can also turn them off or delete them. >> is that shared with outside parties? >> it is not. >> it is only used by twitter? >> yeah. >> when it comes to questions of i would likegment, to think on your thoughts on publisher liability. some have said there is an amount of editorial judgment being carried out when twitter uses artificial intelligence driven algorithms or promotes content through twitter moments. line onould we draw the
1:50 am
how much editorial judgment can be exercised by the owner of a neutral platform before the platform is considered a publisher? >> we do defend section 230 as it is a thing that enables us to increase the health in the first place. it enables us to look at the content, look for abuse, and take enforcement actions against them accordingly. we do have a section of a service called moments, we do have curators who are looking for -- through all of the relevant tweets for a particular event or topic and arranging them. they use an internal guideline to make sure that we are representative of his many perspectives as possible. going back to the concept of variety of perspective. viewnt to see a balanced of what people think about a particular issue. not all of them will be as balanced as others. that's how they measure themselves.
1:51 am
it is one area that people can choose to use or ignore altogether. say thaty, some people twitter could be classified as a media outlet due to certain content agreements. do you have any comment on that? think the broader categories are necessarily useful. servinge our role as conversation. we see our product is a conversational product. a communication product. we see a lot of people use twitter to get the news is we believe that news is a byproduct of public conversation. it allows to see a much broader view of what is currently happening. that is what we are focusing on. how do people use us? we do have partnerships where we
1:52 am
one, itvents like this is live on twitter right now, where people can have a conversation and everyone can benefit. >> thank you. my time is expired. >> the chair recognizes the gentleman from massachusetts. >> thank you so much for being here. thank you for your patience. i know you were in the center side -- senate side earlier today. make sure we were clear on a couple of things. you talked about the mechanism to look ate different aspects of content on the site. you of also talked about how your algorithms are a bit imperfect. how they have impacted some members of this body. democrats and republicans. is that true? >> yes. >> you have also indicated that
1:53 am
there were others who get caught up in it. liberal activists who use profane language. is that true? >> that may or may not be a signal that we use in terms of the content. we tend to favor the behavior we are seeing. that is what i was describing in terms of the signal. the behavior of the people following these accounts. >> fair enough. you yourself were suspended at a time. >> i was. >> fair to say that sometimes there are errors. obstructivengage in dave you're on your own site. which he did not, right? >> correct. automatedked about tools and individual users to report tweets, behavior. that is self-regulation
1:54 am
mechanisms that you all use. >> our motto currently depends on reports to remove content or accounts. >> why do you depend on those reports rather than having a more robust network within your company to do that? what you outsource that users -- why do you outsource that to users? >> we don't believe the burden should be on the victim. this is something we would like to change. we have to build that technology. >> i understand you don't feel great about putting that on the victims are observers. you also express a reticence for your company to be the arbiter as to what is decent, fair, truth. you mentioned the term false fact. i have no idea what that is. toseems like you are trying meld this world of outside crowdsourcing, what works versus internalizing some of it. i want to push you on that.
1:55 am
you say you are trying to fix it. what does that look like? >> we are trying to build a proactive system that recognizes behaviors that are against our terms of service and take action much faster so that people don't have to report them. one of my republican colleagues asked earlier, how many folks you have working on that. you said the issue was and how many people. -- wasn't how many people. i'm certain the technology can advance here. is that two people? 20 people? will you be hiring mark? -- more? >> we have hundreds of people working on it. the reason i don't want to focus on the number is because we need to have the flexibility to make decisions between investing to build more new technology or hiring people. in my experience, companies
1:56 am
naturally want to grow. that isn't always the right answer. it does not allow for a lot of skill building. -- scale building. >> i yield. >> we recognize the gentleman from illinois. >> thank you. thank you for coming in here. i think it is important to mention that twitter has been liberating oppressed people and allowing oppressed people to communicate. if you look into syria, people have been able to get their message out. when chemical weapons attacks happen, we know about that quickly because government censored media, which would never report a chemical weapons attack, is usurped by twitter use. part of a very big concern with that, foreign interference in our democracy.
1:57 am
we are very concerned. i'm very concerned about the role that the russians played in attempting to undermine democracy. i don't think russia elected president trump. it is obvious they are trying to sow instability. i think that the more we can get a grip on this and make people aware of the fact of what is happening, we can begin to an oculi ourselves. -- in oculi ourselves. on september 1, a russian federal law number known as the requires social media companies offering service to russian citizens to collect and maintain all personal information of those citizens on databases physically located in their country. is twitter in compliance with this law? >> i need to follow up with you
1:58 am
on that. >> you don't know? law again?gin -- informationes data to be kept in russia. this has been in the news for a couple of years now. >> i don't. i need to have my team follow up. >> you have a bunch of people back there. you can ask them. >> we don't have servers in russia. >> you are not technically in compliance, which i think is good. that answers my second question. if you store user data, there would be concern about leeches. -- breaches. does twitter make any user data available to russia's state enemies? >> no. >> let me ask you then.
1:59 am
we have touched on this a few times. young adults, teenagers using twitter. i think our laws having caught up with the new reality of the 21st century. we have to address how technology can be used to hurt innocent people. are laws to there prevent people from spreading malicious images. an account can be created in minutes to slander someone. thatuckerberg testified facebook is responsible for the content on facebook. you can appreciate how newsworthy that was given the interpretations of section 230. your agreement clearly states that all content is the sole responsibility of the person who originated it. you may not monitor or control the content. we cannot take responsibility for the content. statements you have made seem to be in conflict with the language. can you take a little bit of time to clarify your stance?
2:00 am
>> in what regard? >> our users responsible? is twitter? is it next -- mixed? >> people are responsible for their content. we have made our singular objective to help improve the health of the content that we see on the service. for us, that means people are not using content to silence others or to harass others or to believe each other so that they don't even feel safe to prevent -- participate in the first post. that is what 230 protects us to do. to actually enforce these actions, make them clear to people in our terms of service. but also to enforce them. >> i'm out of time so i yield. >> the chair recognizes the gentleman from california. >> thank you very much.
2:01 am
thank you for discussing this important matter. i want to follow up. president and republicans are criticizing social media, there are real issues such as the shocking bullied, teens being physical bullying is bad enough but this cruelty is moving online where one click of a button sends a hateful word and images that can be seen by hundreds or even thousands of people at a time. people, kids, are being targeted for being who they are or for being a certain race or sexual orientation. we know it is pervasive. the first lady has made combating cyber bullying a national priority. oddly enough. at the same time, adults are not toing kids a great example follow. public figures including the president speak harmful words every day.
2:02 am
these actions cannot be erased and may follow their victims and families forever. how did it feel to be in front of us for hours at a time? >> i'm enjoying the conversation. >> you get to go home, you get to do what you choose to do once you leave this room? >> yes. >> that is what is incredibly important for us to think about when we think about bullying online. it is inescapable. that is an issue that is new to us as human beings. it can take many forms. it can be hurtful. it is about words and appearances. i think it is really important that the public understands that something needs to be done about it. what can be done is something that we can come to terms with with you over a twitter. on twitter. -- alebrities such as the 40 --
2:03 am
14-year-old from -- have started using -- stopped using twitter because the intensified bullying that they experience on the platform had persisted. help thesewould not public figures, how does it deal with all the kids who aren't famous? handleto know how you bullying claims for american families who are not in the news. love explained that twitter investigates -- you have explained that twitter investigates when it gets reports of bayview like that. -- behavior like that. how many reports of cyber bullying does twitter receive each month? >> we don't disclose that data. we can follow up with you. how about periscope? >> the same. >> ok. look forward to that answer submitted to the committee.
2:04 am
have many of those reports were accounts of people 18 or younger? >> in what regard? you ever take into account whether or not it's a report to somebody who has been attacked with 18 years or younger -- who is 18 years or younger? >> we don't have the same sort of demographic data that are peers do --. r peers do because we are not a service of profiles. >> that makes it more critical for us to understand that. what actions are taken response to these reports? how long does it take? >> we rank according to the severity of the report. again, this is something that we need to improve to understand the severity of. how that is ranked so that we can move much faster. ultimately, we don't want the reporting burden to be on the
2:05 am
victim. we want to do it automatically. >> i'm out of time. think you very much. >> we now turn to the gentleman from virginia. >> thank you very much. i appreciate you being here. i represent that portion of virginia in the southwest corner and borders a big chunk of southern west virginia. i have some questions similar to mr. mckinley. we are suffering from a huge opioid problem. i know that you are trying and are working on it. year, scientific american talked about having artificial intelligence scan twitter for signs of opioid debris is. it would seem to me. on contact, is an illegal that is not just an inconvenience. trying to judge whether it is something that is truly bad. it is illegal.
2:06 am
it seems that you want to be able to employ artificial intelligence that would knock down anybody trying to sell illegal substances on your platform. can you address that? to prioritize all of our models. >> shouldn't illegal be at the top? >> absolutely. we have been prioritizing a lot of what we saw in 2016 in terms of election interference. and our readiness for 2018. here is what i got. i have people writing me whose kids have died because they have been in treatment, they have relapsed, one of the west -- easiest ways to get in there is to get on social media. can use artificial intelligence to track opioid abuse in this country, you want to be able to track illegal sales. wouldn't you agree with that? >> i agree with that. it is horrible.
2:07 am
it is something we need to address. >> i don't think there's a conspiracy. i think that there's a lot of folks out there that may not have that many conservative be living inmight your neighborhood or the area that you live in. i look at your advisory council. there might be some right-leaning groups but i didn't see any rights groups in there. we are not all crazy on the right. find some groups that can help out on your advisory council. i said this to mr. zuckerberg when he was here. it seems to me that if you don't want the government in their, it is better not to have a you wantt in there, to come up with something as a group. 1894 had this newfangled thing. electronic devices were coming onto the scene. an engineer said, we ought to test all of this.
2:08 am
they got the insurance companies and the electric manufacturers together and they funded united laboratories. as an industry, without the government coming in. they came up with standards. it would seem to me that the social media, the big actors, ought to come together, figure out something that is a template that works for all to make sure that we are not having political bias. i really do believe you when you say that you are not trying to do it. it is happening anyway. i think it is an accident. you have to help us. thinkt dig it is good -- it is good for the internet to have the government laying down rules that may or may not make sense. somebody has to do something. we need to protect privacy. there was make sure no political bias, intentional or unintentional. >> that is a great idea. that is why we want to be a lot more open around these health
2:09 am
indicators we are developing. >> one of the questions that do put thep, if you cabal on somebody's post or twitter account, and you tell them about it so they can do the appeal? if they don't know about it, they will not appeal. robusteed a much more way of communicating what happened and why. and on a much more robust process. >> i yelled back. >> turnout to the gentleman from california for four minutes. >> thank you for being here. i don't know if anybody has mentioned the irony that donald trump is complaining about imagine it is hard to that he would've done as well as he did without the platform. he is a master of using it. i think it has done wonderful things for democracy. it has democratized democracy in many ways. we saw that in the house when we
2:10 am
live streamed the protest over guns. it brought people into the chamber in a way that none of us had imagined before. i use it a lot to stay connected back home in san diego. i find out what is going on every day in the local government. i follow my baseball team's promising minor leaks through it. -- leagues through it. now everyone is a journalist. yourt to explore discussion of the term fairness. have you ever written down what you mean by fairness? you have these allegations about fact versus false equivalency. trying to provide truth rather than balance. is that something that goes into your calculation of fairness? what standards do you impose on
2:11 am
content on twitter? , fairness to us means that we are driving more impartial outcomes. more objective driven, not basing anything on bias. we do want to be able to measure this and make public what we find. that is why we kicked off this initiative to understand the health of conversation. and how it might trend. indicators that we are considering is shared fact. that is the percentage of conversation that shares the same facts. that is not an indication of truth or not. what percentage of people participating in a conversation are sharing the same facts versus having different facts? we think a greater collection of shared fact leads to a healthier
2:12 am
conversation. if we understand the makeup of them currently, how can we help drive more people towards sharing more facts? that, we can see a lot more healthy conversations. that is our intent. we are at the phase where we just need to measure it. against those for indicators i laid out earlier. we can send you more of our information, thinking about how we are developing these. >> one of the problems with everything having the wrong fact that it is hard to have conversation -- their own facts, it is hard to have a conversation. one example i have is climate change. herring -- of this hearing be a way to work the refs. is that something we should be concerned about? is that something that strikes you as could have an impact on your business? >> i honestly don't know what
2:13 am
that means. >> the idea is that they're going to put so much pressure on you to avoid pressure from us that he will change her behavior in a way that is not fair. is that something we should be concerned about? >> we have articulated what we think is important. what we are trying to drive. of government as being a checkpoint in that. , --questions of our past path i do believe the system is working in that regard. we are putting out we believe is critical for us to focus on. if there are disagreements and mass, we will certainly change our path. >> my time is expired. >> the chair recognizes the gentleman from florida. >> thank you very much. thank you for your testimony.
2:14 am
local countyom my school district in the west , it has florida consistently responded to threats of school violence. i have heard from the superintendent, he was doing an outstanding job. he faces as many as 19 threats in one week. many of those threats have come from individual tweets. and studies show this is a widespread problem. what is your company's process for notifying local law enforcement officials and school districts when the threats emerge? >> we do have outreach to local entities and local law enforcement when we see anything impacting someone's physical security. we can follow up with you one exactly what those >> limitations are. -- those limitations are. >> how effective have they been? >> i'm not sure how to determine that.
2:15 am
we can follow up with you on that. >> please do. would you consider an internal process in which twitter can work directly with the school districts to address these tweets weekly? time is of the essence. >> yeah. one of the things we are always quickly,or is ways to especially where it impacts physical security, to alert us to things that we might be able to help with in terms of the conversation around it. we are certainly open to it. and open to an implementation we think we could scale. i how did you determine -- know that social media, facebook two, the minimum age of use of 13. are you considering raising that age? >> i don't believe we have considered raising the age. we did determine it. >> thank you.
2:16 am
website, to twitter's ,witter moments are defined as quote, curated stories showing the very best of what is happening on twitter. and customized to show you topics that are popular or relevant so that you can discover what is unfolding, customized to show you topics. in my experience, twitter moments more often feature a specific point of view or political narrative. compiledhese moments and prioritized? you said earlier that moments are selected by employees publishing content. what are the internal guidelines the company has set to determine what makes a moment?
2:17 am
>> we first and foremost take a data driven approach to how we arrange these moments. these are collections of tweets that we look at aced on any particular topic or event. -- based on any particular topic or event. we bring them into a collection. driven approach, meaning that we are looking for the amount of conversation that is happening around a particular event. as we rank that, we go into impartiality to make sure that we are looking for opportunities to show as many perspectives as possible. a variety of perspectives and a to theore is beneficial people reading because they can see every side of a particular issue or event. very good. i thank you and look forward to
2:18 am
getting information, following up. we would like to have you very . back, even though i'm not the chairman. to see the progress that you have made with regard to these issues. i yield back. >> the chair recognizes the young lady from michigan. >> thank you. you are one of my husband's heroes. to the dean of twitter, who is better on twitter than anybody in this room. power ofs i know the this platform. i think it is very -- a very important tool. for those who have been politicizing this, it is not only meghan mccain -- i have had some of those threats. i was blissfully ignorant until law enforcement brought it to my attention. that the threats that
2:19 am
are happening on twitter need to be better understood and more quickly acted upon. i would rather ask questions right now. you are educating all of us and we all need to understand social media better. i would like to ask questions about privacy. and the use of machine learning and artificial intelligence on the platform. you'll spoken about how you are trying to deploy machine learning to combat the disinformation, the harassment, the abuse. i want to build on what some of my other colleagues have said about the blackbox nature of these algorithms and the lack of what they call accountability. colleaguen what my representative harper was saying, what kind of data sets do you use to train i high? -- ai? how often do you retrain them? >> we try to use datasets that will be predictive of what we
2:20 am
would expect to see on the service. as we train these models, we are experiences and outputs that we have seen in natural uses of how people use the system. testhen also trying to some edge cases as well. great andse tests are help us understand what to expect area ultimately, they are not but the test until they are released on production. that is training. there are ai, the application of ai at scale is relatively new. especially to us in our company. there are best practices being developed that we are learning as quickly as possible from. more importantly, trying to
2:21 am
measure those outcomes in terms of bias and impartiality. >> do your engineers have an ability to see and understand why an algorithm made certain decisions? >> that is a great question. that goes into another field of research in ai which is called explain ability. encouraging engineers to write a function that enables the algorithm to describe how it made the decision and why it made the decision. i think that is a critical question to ask and focus on. we are offloading more not -- more and more of our decisions to these technologies. even personally. i am wearing an apple watch right now and it tells me when to stand. i have offloaded a decision to that.
2:22 am
it cannot explain why it is taking that action. it became's quite -- that comes quite scary. i believe that is a valid form but it is extremely early in terms of research. bility,ncept of explaina . >> i will be out of time. havealked about how these missed things, it has made mistakes. what is an acceptable error? you can do that for the record later. >> the chair now recognizes the gentleman from ohio. for being here today. say that ano algorithm is essentially a , once it has that turned into software, it operates on a data set and
2:23 am
produces a desired result? would that be a good layman's term? >> for a general algorithm, yes. it gets a lot more,. >> i will get into the complication. i'm a software engineer by trade. i have written thousands of algorithms. there's as much art that goes into writing an algorithm as the wrist science. -- as there is science. would you agree with that? >> yes. >> there is a part of the heart of the algorithm writer writing the algorithm. correct? painter, if you put 10 painters in 10 different different will get paintings in each room. if you asked 10 software engineers to develop an algorithm, you will get 10 different solutions. right?
2:24 am
>> that is why testing is so important. >> what kind appeared testing to you do -- do you do with your algorithms to make sure that that innate bias -- that is inevitable. it has been admitted that twitter has bias in its algorithms. you have acknowledged that and try to correct it. how do you go about weeding that bias? do you do peer-reviewed's before you send them to production? >> we do those internally, yes. modify your be morems to intelligent and i -- in identifying specific things? in the automotive industry today, we have artificial intelligence that doesn't just tell you that there's a car in front of you, it actually puts
2:25 am
the brakes on. it takes action. it is instantaneous. it saves lives. is it unreasonable to think that twitter could not modify its algorithms to hit on illegal drug sales? on violent terminology? those kinds of things. and make faster alerts to stop some of this. >> not unreasonable at all. it is a matter of work. >> you need to do that work. you need to get to a quick. the trending topics list is an important issue. i want to understand that one. topic isell me how a determined to be trending? give me specifics. what is a based on? >> in a tweet, when you use a particular keyword or h
2:26 am
ashtag, the system recognizes when they are used in mass. it recognizes when there is a velocity shift in the number of times people are tweeting about a particular trend. it identifies those and puts them in the trending topic list. there is a default setting or we personalize those trending topics for you. that is a default. when you first come onto twitter , trending topics are personalized to you. that is based on the accounts you follow anti-you engage with tweets and whatnot. we could show you all the trending topics happening in the world but not all of them are going to be relevant to you. we take the ones that are in drink them accordingly. >> it is trending based on what is relevant to you. >> correct. >> let me just say this. i said this to mr. zuckerberg. in the absence of massive federal regulations telling you guys how to do your business, the responsibility bar is really
2:27 am
high. coming back to what mr. griffith says, you need to look at an sort toentity of some help you bounce off ideas of how to address this stuff before legal or market forces drive you to a place that you are not going to want to go. >> gentlemen, tie. the chair recognizes the gentleman from new york. >> thank you for all the time you have given the committee. i want to echo my dismay that our republican colleagues have chosen to pull this hearing to rile up their base and give credence to unsupported conspiracies when there are real issues here that run to the heart of our civic life that deserve our immediate attention. missedn unfortunate opportunity. has said ittwitter
2:28 am
is taking steps to help make political advertising more transparent. it is now working to do something similar with issue ads. today, i amwitter concerned that even for political ads, you haven't made anything clearer to consumers. some platforms, facebook as an example, if a user visits a website, or politician that user can immediately see all the advertisements that they have purchased. on twitter, i have to find a separate resource, the ads transparency center, and search the politicians to see what promotions they purchased in the past. to,eems a little eyes particularly -- ill advised, regularly -- particularly when you're peers are doing it
2:29 am
better. we did do some research. this is not a stopping point for us. we want to continue to make transparency something that is meeting our customers were they are. so that it is relevant. so that it is easy to get to. we did some things a little bit differently. we have launched the issue ads feature of the transference agency as well. search enabled anyone to twitter ads to see who was behind them. and also the targeting criteria that are used. statistics kept any they can share with this committee today about how often the average consumers even search the adds transparency center? >> we do keep statistics and track usage of all of our products.
2:30 am
we can certainly follow up with your office to give you some relevant information. >> thank you. i know that you said this is not a stopping point, you are still exploring. why is that it appears you're making it harder for americans to see who is trying to influence them? mr. dorsey: that's not our intention. and we do need -- we know we need to make -- to do a lot more work to meet people where they are, there's some design choices we need to make in order to do this the right way. mr. tonko: it seems political advertising information that twitter makes available only shows advertisements served in the past seven day, is that correct? mr. dorsey: i'm not aware of the constraints on it but we'll follow up with you. mr. tonko: if that is correct it seems vastly insufficient given that political campaigns in the u.s. last months if not years. mr. dorsey, why doesn't your platform reflect that insight and disclose political advertising beyond seven days if
2:31 am
that indeed is the limit. mr. dorsey: we'll look into that. mr. tonko: thank you, appreciate that. i yield back the balance of my time. >> we now go to the gentleman from missouri, mr. long, for four minutes. mr. long: thank you. think it is pretty easy to understand why you've been as successful as you have. your mannerisms and your decorum. a will the of people come to these hearings, people coach them and tell them how to act. it's obvious that's not what happened with you. thank you for your time and being here today. mr. dorsey: thank you. >> i do have a couple of questions. mr. bilirakis asked you about moments, i'm not sure what moments are but when any of my staff got a hold of me a couple of days and said what do you want to ask mr. dorsey and what direction do you want to take this? i did a really research.
2:32 am
-- little research. off the top of my head i said let me send you some stuff. i start shoot them emails. these are emails i received, they're called highlights, you're familiar with, daily highlights to my personal twitter account about the most interesting content from twitter tailored just for me. and when we're talking about impartiality and you know, somebody said the republicans are all fuel of -- all full of conspiracy theories over here, i want you to -- you're a thoughtful guy i want you to take into consideration what i'm going to say and do with it what you want to. but if you're saying hey, we're impartial, we are, this, that and the other, out of the -- i pulled just started firing off emails to my leg director and sent him 14 emails of highlights that were sent to me in the last few days. may have been over 14 days, i don't know how often you send them. there are six highlighted tweets per email. a total of 84 recent examples
2:33 am
that you all picked and said this conservative congressman from missouri, thank goodness you're a cardinal fan and you being from missouri but this conservative congressman we found out what this guy wants to read. here it is. 12 of them of the 84 were in glenn thrush of "the new york imes." maggie, you sent me nine from her, white house correspondent for "the new york times." political analyst for cnn. a political commentator from cnn. david frum, of "the atlantic," nicole wallace, current anchor of "deadline" and report for msnbc. sam stein former political editor of "huffington post." politics editor at daily beast." msnbc contributor. representative eric swalwell democratic congressman from california's 15th district. robert costa, national political reporter for "the washington post," political analyst for nbc news and msnbc. kaitlyn collins, white house correspondent for cnn. michael schmitt, "new york times" correspondent,
2:34 am
contributor to msnbc and nbc. tommy veter, former spokesman for president obama. david korn, msnbcablist and author of the rution lieu roux let book. casey hutton, nbc news correspondent, host of msnbc show. richard painter, commentator on msnbc and cnn outspoken critic of trump. david axelrod, former chief strategist for obama's campaign, senior advisor to obama. i did not cherry pick these. here's a republican you did send me one from bill crystal founder and editor of the at-large weekly and a vocal supporter and never trumper guy. an you did send me one from fox news, i'll put that in there, brit hume, senior political analyst for fox news channel. i want to submit these for the record so you can peruse them at your leisure. that's only two i remember being epublican.
2:35 am
out of 84, hand picked tailored for me because they know what i want to read. glen thrush,, it just goes on and on. i have i guess 14 pages of them here. and they're all pretty much trump bashing. they're all pretty much trump bashing. if you go right down the line, one after another. o just if you will take that into consideration and i think that was fake news tweet sent out yesterday by a guy sitting here earlier and he's not here anymore the reporter sent out a fake news tweet yesterday who said he was headed to the nationals park an he was going to watch them beat the cardinals. that was fake news. i yield back. mr. dorsey: thank you. doesn't sound like we served you well in matching your interests. >> the gentleman's time has expire the chair recognizes ms. schakowsky.
2:36 am
ms. schakowsky: while you've sitting here all day, we appreciate that, the -- according to "the wall street journal," the justice department, to examine whether social media giants are, quote, intentionally stifling, unquote, some viewpoints. and it closed -- it quotes the president, it says in an interview wednesday morning with the daily caller, mr. trump accused social media companies of interfering in elections in favor of democrats, quote, the truth is, they were all on hillary clinton's side, he aid. would you agree with that characterization by the president? mr. dorsey: no. ms. schakowsky: the other thing it says in this article is that they expressed, referring to -- the -- i guess it's in the senate -- the -- they expressed contrition for allowing their
2:37 am
platform to be abused in the past while pledging to make protecting the system from -- protecting the system from during the 2018 mid-term elections a priority. first of all, i just want to say about contrition, we heard from facebook c.e.o., mr. zuckerberg, one example after another after another through the years, you haven't been there that long at twitter, but -- of contrition. we're sorry, we're sorry, we're sorry. but even today, if i had listed, well we made a mistake, we're going to do better, etc. so first let me ask you, what are you going to do to make sure that the election is not in some way influenced by foreign governments? in an inappropriate way? mr. dorsey: that our number one priority and our information quality efforts. ms. schakowsky: i hear that.
2:38 am
mr. dorsey: we have benefited from learning from recent elections like the mexican election and we're able to test and refine a bunch of that work accordingly. we're doing a few things. we're opening portals that allow partners and journalists to report anything suspicious that they see so we can take much faster action. second, we are utilizing more technology to identify where people are trying to artificially amplify information to steer or detract the conversation. third, we have a much stronger partnership with law enforcement, federal law enforcement to make sure that we are getting a regular cadence of meetings that we are seeing more of the trends going on and that we can understand intent behind these accounts and activities so we can act much faster as well. ms. schakowsky: i appreciate that. that's where the emphasis ought to be. i have to tell you the president and the republicans have
2:39 am
concocted this idea of a supposed anti-conservative bias to, it seems to me, distract from the fact that the -- that their majority has absolutely done nothing to prevent foreign governments from using social media platforms to spread misinformation. and if we don't do that, then i think our democracy itself is actually at stake. but also in terms of your motives, mr. dorsey, the majority of twitter's revenue comes from selling advertising on the platform, right? mr. dorsey: right. ms. schakowsky: and it's a publicly traded company, right? mr. dorsey: correct. ms. schakowsky: and campaigns choose to advertise on twitter because twitter promises to deliver targeted, highly engaged audience, is that what you'd say? mr. dorsey: right. ms. schakowsky: so you said you
2:40 am
are incentivized, it says twitter is incentivized to keep all voices on the platform is that correct? mr. dorsey: no, that is where we need to make sure we're questioning our incentives but also we understand that making health our top and singular priority means that we are going to be removing accounts and we have done that. ms. schakowsky: i am quoting that you said from a business perspective twitter is incentivized to keep all voices on the platform. mr. dorsey: all perspectives. but i thought you meant more the accounts. we want to make sure that we are -- we believe we're used as a public square. for people. and that all perspectives should be represented. ms. schakowsky: thank you, and thank you for being here. >> the gentlelady's time has expired. the chair recognizes the gentleman from indiana, mr. bucshon. mr. bucshon: i don't see this as particularly partisan, the hearing.
2:41 am
i think it's completely appropriate and relevant to the american people across political ideologies. i would respectfully disagree with my democratic colleagues and some of the colleagues they've made. i would like to say this, ironically in my view, they're the ones most likely to want heavy handed government intervention into your industry and i would argue people like me, republicans, are trying to help you avoid it. so take that for what it's worth. you know, you have implied and you've said that twitter is taking all these different actions to improve all things that you're doing as it relates to algorithms, what's your timeline? and i know you've said this is an ongoing process, you're never going -- you're not going to get a checkered flag, but what's your timeline for getting some of this really done? mr. dorsey: we want to move as fast as possible and i know that's a frustrating answer because it's hard to predict these outcomes and how long they
2:42 am
may take but it is our singular objective as a company in terms of improving the -- increasing the health of the public square we're hosting. mr. bucshon: thank you. is there any way users and third parties can verify whether or not their political standards or judgments are embedded accidently into twitter's algorithms? i guess i'm asking, are your algorithms publicly available for independent coders to assess whether there is bias, whether it's intended or unintended? mr. dorsey: not today but that is an area we are looking at and we'd love to be more open as a company, including our algorithms and how they work. we don't yet know the best way to do that. we also have to consider in some cases when we are more clear about how our algorithms work it allows for gaming of the system so people taking advantage of
2:43 am
it. so we need to be cognizant of that. it's not a blocker by any mean, e'd love for it to be open but that's a big understanding we need to understand how to correct. mr. wueshon: i understand that. if algorithms are there smart peep will find ways to subvert that and there's probably some proprietaryness there, you may have a competitor in the future ou don't want your processes out this. i totally respect that. mr. dorsey: this is an area we do not want to compete on health we want to share whatever we find. mr. bucshon: i think many people have said, all of us, whether we know it or not have some inherent biases based on where we grew up, what you are background is what our life experiences are. i'm kind of -- i'm interested in how you recruit, you know, to your company. because i think, obviously the tech industry has had some
2:44 am
criticism about its level of diversity but i think it would be important to get your feel for, if you're going to avoid groupthink and you're creating algorithms, how do you recruit? you're not going to ask somebody, are you pro trumpe or or against trump? i get that, right? but i would argue you want to have people from everywhere, different race, men, women, different political views. because our -- my impression is diversity is in some respects in certain industries fine as long as it's not political diversity. so how do you, can you give me a sense of how you build a team? mr. dorsey: this is an active conversation within the company right now. we recognize that we need to decentralize our workforce out of san francisco. not everyone wants to be in san francisco, not everyone wants to work in san francisco, not everyone can afford to even come close to living in san francisco. and it's not fair. so we're considering ways of how
2:45 am
we hire more broadly across every geography, across this country and around the world. and being a lot more flexible. it's finally the case that technology is enabling more of that. and we're really excited about this. i'm personally excited to not consider san francisco to be a headquarter bus to be a more distributed company. mr. bucshon: i think it's important to make sure that companies like yours do get a variety of perspectives within your employee base. thank you. thanks for being here. mr. dorsey: thank you. >> the chair recognizes the gentleman from california, for four minutes. >> mr. dorsey. had a long day. you're in the homestretch. i'm glad my colleagues on this side of the aisle have focusing on the issues important to our democracy, how we combat foreign influences, thoughts, harassment and other challenges on the platform. i'd like to take a step back and look at the makeup of twitter's users.
2:46 am
mr. ruiz: i'm not we or you have a true understanding of who is really using your services and your website. because you have said previously the numb of followers an account has is critically important in terms of the prominence of an account but also the ranking of algorithms that push content to users. so with tens of thousands of new accounts created every day, real nd fake, by humans and bots alike, i'm concerned about the accuracy of the numbers we are using here today and the implications those numbers have. you have said that 5% of your accounts are false or spam accounts is is that correct? mr. dorsey: correct. mr. ruiz: is that at -- how do you measure that? is that at any one time or over the course of any one year, how did you come to the conclusion of 5%? mr. dorsey: we have various methods of identification, most of them automations and machine learning algorithms to identify these in realtime look thought
2:47 am
may havors of those accounts. mr. ruiz: that's how you determine which are false but how did you come up with the 5% estimate of total users are fake? mr. dorsey: it's 5% we believe are taking on spammy like behaviors which would indicate automation or some sort of coordination to amplify information beyond their earned reach. so we're looking at behaviors. mr. ruiz: but that number versus -- mr. dorsey: the total active. that number has remained fairly consistent over time. mr. ruiz: in 2015 you reported 302 million active monthly users on your platform. in 2016, 317 million monthly active users. 2017, 330 million, and then in 018 you said 335 million monthly active users. how do you define that? mr. dorsey: someone who engages with the service in a month.
2:48 am
mr. ruiz: someone who tweet, retweets or logs in? mr. dorsey: someone who logs in. mr. ruiz: and it's 5% of that number that you believe to be spam, of speedway who just logs in? mr. dorsey: yes. and taking on spam behaviors. mr. ruiz: and has the 5% been consistent over the years? mr. dorsey: it has been consistent. mr. ruiz: we have heard reports of hundreds of twitter accounts run by one person. it's my understanding that each of those is counted as separate monthly active users, is that correct? mr. dorsey: correct. mr. ruiz: my concern is the number of followers an account has, comprised of the subset of those 335 million twitter users is an incredibly important metric to your site and one you said before the the senate, presents too much incentive for account holders. based on what we've heard it appears the number of followers
2:49 am
may not be an accurate representation of how many real people follow any given account. for example, last year twitter added roughly 13 million users but early today you said you are flagging or removing 8 million to 10 million per week. how can we be confident the 5% fraudulent account number you are citing is accurate? mr. dorsey: we're constantly updating our numbers and understanding of our system and getting better and better at that. we do see our work to mitigate mr. ruiz: before we end time, i'll ask one question, you can submit the information, that's basically in medicine or any screening utility, i'm a doctor for any screening utility we use a specificity and sensitivity and that measures how well your methodology works. and the higher specificity, the lower false positive you have. the higher sensitivity, the
2:50 am
lower false negatives that you have. in this case, you can see how many false negatives versus false positives. we're concerned you're going to have false negatives with the russian bots. and that you're taking out people that should be on there. if you can report to us what those specificity is in identifying bots, i would appreciate that. that would give us a sense where your strengths and weaknesses are. >> time expired. >> thank you, mr. chair. i appreciate you showing up today. if you don't mind i'm going to round up a bunch of questions and ask to answer those later and then i have a question or two at the close that i would like to try to get asked. our local broadcasters provide a valuable service when it comes
2:51 am
to broadcasting at different events that happened. talkeard mr. burjs earlier about the tv station and the attack. my question is this. should twitter be a trusted emergency alerting system and how do you hang in spread of misinformation or abuse and you n accept mentally -- supplementally answer that if you would. the next question has to do with free speech. determining whether a user has violated its rules or is it only once another user voices the concerns? the next question is do you have a set of values that twitter follows when it makes the decisions flagged content or is it only a case by case basis and
2:52 am
have individuals at twitter make judgment calls? the next one has to do with -- how do you -- this is a conceptual question. how do you balance filtering and moderating free speech? there is this tenuous balance. if you could, i would like to have you respond to that. this is an oversight hearing. we're trying to legislate. we're not trying to get into fights. we're trying to learn about the space. what we would like to have are twitter's definitions of behavior and hateful conduct. twitter's definition of low quality tweets. i would like an explanation of the abuse reports process and lso you said you had signals for ranking and filtering. i would like to know how that works.
2:53 am
trust and safety and how that works. some of that is publicly available. some of it is not. and then twitter definition of suspicious activity. here is the question i have in the last minute i would like you to respond to. a lot of it has been over tumultuous times. my question is this. if we were to have a hearing a year from now, what would be the three biggest changes that twitter has made that you would share with congress? mr. dorsey: that is an excellent question. i believe first and foremost we see a lot of progress on increasing the health of public conversation. second, i believe i be that we have re-- believe that we have reduced much of the burden a victim has to go through to report any content that is against them or silencing their voice or causing them to not want to participate in the public space in the first place. deepern third, we have a
2:54 am
understanding of the real world effects of platform of our service, both to the broader public and also to the individual as well. and those are things that i think we can land make a lot of progress on. the latter one being probably the hardest to determine, but i think we're going to learn a lot within these 2018 elections. >> ok. i thank you for your responses. i know you have 10 people back there that took good notes on the ones i left. thank you. >> yields back. the gentleman from illinois, mr. rush is recognized for four minutes. >> i want to thank you for being here and for enduring this marathon of questions.
2:55 am
i'm going to go back to the --inning of this hearing and expressed the need an independent third party, consumer rights on twitter. i'm not sure of your answer. it was kind of vague to me. i have a question. or you willing to commit to are you saying that twitter will consider the request? is that a commitment or is it just a consideration? mr. dorsey: we are willing to commit to working with you and staff to understand how to do this best. in a way that is actually going to show what we can track and the results. but i think that is the dialogue
2:56 am
we need to have. >> thank you. chicago is experiencing an epidemic of violence, especially relating to our young people. facebook has already been -- by somes an asset of these young people to commit violence purchasing my question if twitter was used to organize any form of violence in the nation? mr. dorsey: we do look at cases and report where is people are utilizing twitter and coordinating in term terms of having platform violence. we have an extremist group where e look at off-platform
2:57 am
information to make judgments. >> is there a proosess for removal of such posts? mr. dorsey: yes, sir. there is a reporting process but it requires right now for removal of post a report of the violation. > how are they removed though? mr. dorsey: how many have been removed? >> have you removed any? mr. dorsey: we do often remove content that violates our terms of service. we have a series of enforcement actions that ranges from a to temporary suspension and removal on a tweet to permanent suspension of the account. -- you also have any
2:58 am
olice department of these kind of activities? mr. dorsey: we do have partnerships with local enforcement and law enforcement agencies all over the world and we do inform them as necessary. >> let me ask you one final question here. for a defendant or anyone else when it comes to abusive tweeting. do you consider president trump's tweets to be abusive at all? mr. dorsey: we hold every account to the same standards and consistency of our enforcement. we do have a clause within our terms of service that allows for public interest and
2:59 am
understanding of public interest per tweet and we definitely weigh that as we consider enforcement. >> mr. chairman, i seek -- on behalf of our colleague from california. >> without objection. the gentlelady from indiana is recognized for four minutes. >> thank you for sitting here today for a very long day and a lot of questions. i want to share with you a little bit on the public safety angle. in 2015 i was pleased because we got signed into law the department of homeland security, social media improvement acts bill. this group has been meeting.
3:00 am
they have issued about three different reports and i actually one of the reports focus on highlighting emergencies. is on social media in public safety exercises. and then another is how they operationalize social media for public safety. curious if your team anything about this group and whether or not your team might be willing to assist this group. while i recognize you have contacts around the globe, there , a safetys a group group, that is focused on this, and we need better interaction with platforms and public safety communities. is that something you might be willing to consider? it, but not aware of
3:01 am
i'm sure my team will consider a -- it. >> i am curious and i asked mr. zuckerberg this, with respect to terrorism and extremist groups down,ou monitor and take i have seen reports that you have taken down 274,000 in a six month period. largeeems like a very number of accounts. and people believe it is not happening. i understand you have worked with others to create a shared database. is it still in use?
3:02 am
>> we are still working together, and this is an active collaboration. a lot of the work we have been doing continues to bear fruit. we are happy to send the committee more detailed results. ask, you have to talked about the report, it is not done yet, right? >> it is not finish for actions upon content and accounts that have to do with our health aspects. all of these questions that you have gotten, and there have been a lot, can we expect that these things might be in that report? the first step is to figure out what is most meaningful to put in their. -- there. designing the document so that
3:03 am
people can get meaningful input. and then we need to aggregate the data. we are in the early phases of designing the document. but we would like to move fast, because we believe it would help trust. a public safety perspective, you cannot and should not dissolve -- and i voluntary do. i appreciate it is important to have an important dialogue and to have as much information as possible in the conversation, i certainly hope that your work with law enforcement, we need to make sure the bad guys don't understand what you are doing. so i thank you and look forward to your continued work. >> the chair now the gentleman from pennsylvania. in your testimony, you identify a handful of behavioral
3:04 am
signals and notes that twitter uses thousands of signals in your behavioral ranking models. can you provide the committee with a complete accounting of these signals? >> a loss of those signals are changing constantly. so even if we percent want today, it might change within a week. not a thousand behavioral signals, it is 1000 decision-making criteria for the algorithms use. i do not mean exactly 1000. they all very. -- vary. providingou consider an expansive list of signals? especially those that indoor? -- endure? we do not have conclusions just yet. the reason we are causing here
3:05 am
in considering is that, by giving up certain criteria, we may be enabling more gaming of the system. >> you use the term a little earlier. chair-ators. what is that? product within twitter called "moments. " if you go to the search icon, you can see a collection of tweets arranged by human, organize around a particular event. it might be a sporting game, for example. to look forcurators all of the tweets that would be relevant. and one of the things he wants to ensure is that we are seeing different perspectives. >> is that relevant space on my behavior? we do the work, and sometimes
3:06 am
you may get a moment but is more personalized. in some cases, all people get the same moment. subject -- would that be subject and open up consideration for more bias, in any way. are making some sort of subjective determination on what might be of interest? they use a data driven approach based on the percentage of conversation that people are seeing. we are trying to reflect how much of this is being talked about on the network and then checking it against impartiality and making sure we are increasing the variety of perspective. >> i appreciate your written testimony.
3:07 am
he said something that interested me. one was that you have no incentive to remove people. no incentive to remove conservatives, because the more people talking, the better. but strikes me that, when we are talking about hate speech or personal information, things that are just straight up mean, there is an incentive not to remove that stuff if it is driving more participation. how do you reconcile that? >> excellent question. and something that we have balanced in terms of, one, or similar objective is to increase the health of this public square and the space. in the short term, that will mean removing all caps. --accounts. we do believe that increasing the health on twitter is growth, but only long-term.
3:08 am
over the past few months, we have taken a lot of actions to remove accounts and mess -- and the mass -- en masse. we believe because that, over the long-term, these are the right moves so we may continue to serve a healthy space. >> we now recognize the gentleman from oklahoma. >> thank you chairman and mr. dorsey. i have got a question, and this is not a gotcha question, a point to which i want to make. virginia,eague from mr. griffin, said earlier. we do not believe it is on purpose, it is just the way things are working out, the totem to which you guys use figure out who will be censored and two is not.
3:09 am
-- who is not. would you consider yourself conservative, liberal? how would you consider your political views? >> i try to focus on the issues. >> i know, but these issues are and. -- at >> are you a registered voter? >> independent. myself,business owner different departments i have seem to take on the personality of the ones that i have running it. the people i have running the department or business or organization. when i stepped down as ceo of my company, the new ceo took on a different personality and the employees followed. we are choosing one mindset over
3:10 am
another in some way, regardless if it is on purpose or not. the way it is being picked, portrayed, it is somewhat obvious. let me simply make my point. in 2016, the presidential , twitter was accused of suspending an anti-hillary focused account and deemphasized popular hashtags. twitter part of a campaign video for an advertising platform, calling it inflammatory. 2017, a single rogue employee deactivated trumps account for 11 minutes. a shocking that a single rogue employee could have that much authority. int is a different question july 2018, twitter was accused of limiting visibility of certain republican politicians by preventing their official accounts from appearing inside
3:11 am
auto populated drop-down search bar results. later, a conservative activist account was suspended after, essentially, imitating an account from a new york times editorial board member. are you familiar with this? let me read what they wrote. #cancel white people. white people marking up the internet with their opinions like dogs missing on fire hydrants. our white people genetically predisposed to burn faster in the sun, thus logically being fit to only live underground? enjoy --k how much i or how much joy i get out of the cruel to old white men.
3:12 am
greet ay mouth to republican, but nothing but an unending cascade of vomiting flows from my mouth. out byme tweet went candace owens, but replaced jewish for weight -- white. miss owens account was suspended and flagged. the new york times reporter account was not. what is the difference? >> we did make a mistake. >> i have heard you say that multiple times. i have heard to say that the whole time you have been up. , and you have been very polite and doing it. but the fact is it is bigger. it is the environment that i think twitter has. my point of the first question was, does that fit your political views that your company is following?
3:13 am
because there seems to be a pattern here. it doesn't, i value variety in perspective and seeing people from all walks of life and points of views. errors along the way, both in terms of our algorithms and in terms of the people following guidelines. >> the gentleman's time is expired. >> i wreck of -- recognize the gentleman from michigan. >> thank you. the only complaint i would have is that your staff did not care well enough to go through 535 members of congress. and have those figures for us today. i would assume that they should have thought that come with republicans and democrats here and the statements we have heard, that that question would come up. members, 535he
3:14 am
members, and would have been worth being handled to answer today with an imperative, no there was no bias. that is the only complaint i have. but let me go to the questions. twittery 26 blog post, asserted, and i quote, we believe the issue had more to do with how others were interacting with these representative accounts. what specific signals and suggest?ould you >> the behaviors we were seeing were actual violations of our terms of service. >> would it mutate or blocking another user's account contribute to that? >> note, these were reported violations we had found. >> in retweeting or boosting
3:15 am
would not be a contribution? >> no. does twitter have policies and procedures in place to notify accounts or users when their messages or content have been hidden from other users? >> we do not have enough of this. we do have a lot of work to do to help people understand why we might rank or filter or put their content behind. that is an area of improvement. we have not done enough. while i appreciate you do not want to have users be responsible for contacting you about issues, you want to be catching some of this stuff. or have no specific timeline strong policy in place to notify , me, for instance, that there
3:16 am
is a reason why you had taken me down. so that i am a productive and positive member. >> we take any enforcement action that results in removal, you are notified immediately. if it is a question of filtering , then we do not have a great way of doing this today. it is our intention to look deeper into this. i know this is a frustrating answer, but the timelines are a little unpredictable. thate do believe transparency is an important concept for us to push. we want to earn trust. >> with regard to internet service providers, they are required to disclose if they are throttling or blocking services. would you be open to a similar
3:17 am
set of transparency rules when you have taken actions that could be viewed as blocking or throttling of content? >> we are considering a transparency report around our actions. we are in the faces right now of understanding what is going to be most useful. and then to do the engineering work to put in place. and think it is a good idea something that helps earn trust. >> i wish you well. i do not want to be my colleagues that wants to regulate. this is an amazing opportunity we have. want to see government to get involved in regulating if you folks can do the job results. we recognize mr. duncan. >> thank you chairman and mr. dorsey. we have heard a lot about content filters and a little bit about bias.
3:18 am
i would like to focus on by this. created af my staff test twitter account working on a communications product unrelated to this topic and before we knew this hearing would take place. they were interested to note who was listed on the suggestions for you to follow list. a pro-life conservative congressional staffer of a work computer whose search history does not lean left. all the entered was an email address and a 202 area code phone number. yet here is to twitter suggested they follow. nancy pelosi, kamala harris, chuck schumer, david axelrod, kirsten gillibran, jim acosta. matalin all right, claire mccaskill, choctaw, and john lovett. that is all she got suggested. forget the fact that there are no republicans or conservatives on the list. no singers, no actors, no
3:19 am
athlete, no celebrities. she is a twentysomething female staffer, did not even get taylor swift, chris pratt, cristiano ronaldo, or kim kardashian. all she got the suggestions i had on the street. -- screen. it is one thing to not promote conservatives, even though donald trump is the most successful twitter user. say what you want about how he tweets, but he has utilized twitter in unprecedented ways. i would think that someone in your position would be celebrating that and him rather than trying to undermine it. female,ou explain how a twentysomething year old who put it in an email address and a 202 area code. why do she only gets a liberal suggestion? >> we simply do not have enough information in that case to build up a more informed
3:20 am
suggestion firm. the 202 number is all we have. therefore -- >> i get that you don't have information. 100% of the suggested followers were biased. where was kim kardashian? taylor swift? where was ariana grande? i can look at twitter's most followers, and they are not these people you suggested. and there was nothing in her search history to suggest that she was left-leaning or anything. katy perry? number one. she wasn't on the list. looking at the 202 as a decent number and taking dc-based accounts and the most followed or must engagement accounts. where is price harper, ovitz can, the capitals? where is d.c. united?
3:21 am
where are the sports teams? if you are going to use the 202 area code, where are those folks? there are no athletes, no singers, no celebrities. there were only suggested political figures of a very liberal persuasion that were suggested. nobody else. that shows bias, sir. yeah, i mean, we do have more work to do in terms of our onboard. you are pointing out some weaknesses in our signals that we use. if she were to start following different accounts or engage in with particular tweets, that model would completely change based on that. we just do not have enough information. being like we are not exhausted enough with the one-piece of information we do
3:22 am
have, her area code. >> but me ask you this. and mehis hearing showing this bias, if somebody in a 202 area code was 28 years anti-militant,- are you going to tell me today that they will get other suggested followers than the liberals that i mentioned? >> that is not a good outcome for us. >> we recognize the lady from california, miss walters. thatws indicates periscope, twitter's live video is being used, to sexually target children. activity coordinated is employed to persuade children to engage in sexual behavior. these videos can be live streamed in public or private broadcast on paris to --
3:23 am
periscope. i recognize that a live video app can create challenges, especially when trying to monitor content in real-time. yet your content regarding twitter reads that "we strongly believe that any such activity is unacceptable. " i hope that standard is simply -- similarly applied to content on periscope. is, regarding the twitter has regarded a zero-tolerance policy. does twitter primarily rely on users to report sexually inappropriate content or content concerning child safety? >> we do have some dependency on reports, but this is an area we want to move much faster in automating. and jobs, obviously, not placing the work on the victim. and making sure we are recognizing these and we have
3:24 am
made some progress. >> what is the average length of a live video? >> i am not aware of that, but we can follow up. what is the average response time to remove a live video on periscope? >> it depends entirely on the severity of the report and the context. we try to prioritize by severity. suicidalf death or tendencies would get a higher priority than everything else. >> out of curiosity, when you say we try to eliminate, we have a higher priority, who makes that decision? >> when people report any violations of our terms of we have albertans looking at the report trying to understand how to prioritize those reports to they are seen by humans.
3:25 am
>> i would assume you do not believe that user reporting is an effective method? >> not over the long-term. >> is user reporting ineffective and effective method on periscope? >> not over the long-term. that is something we need to do much more workaround. can you indicate that you need to do more work? do you have a timeframe of when you will be able to handle this? to work as like quickly as possible and make sure we are prioritizing the proactive approaches of our doescement, and again, it go down that prioritization staff. -- staff. i know it is frustrating to not hear a particular time frame,
3:26 am
but we are moving fast. >> can you explain the technology you are using to change this? machinell be using learning and deeply in order to look at all of our systems at the rightprioritize review cadence. >> i yield. the chair recognizes mr. carter from georgia, our last member to participate. thank you mr. chairman, and mr. dorsey, congrats, i am the last one. in preparation of this, i sent out questions to my district and asked them what do you think i ought to ask? i got back some interesting responses, and one came from a teenage high school student, a conservative student down in camden county, right on the
3:27 am
.eorgia florida state line he said i am a conservative high school student, and i on twitter , and i have got over 40,000 followers. tried fiveman had times to get verification. and yet he has been turned down. haveuestion to me was, i got friends who are more liberal than me who have got less followers and yet they have been verified. why is that? what should i tell him? >> first and foremost, we need a complete reboot of our verification system. it is not serving us or the people well. it really depends on when his friends were verified. verificationen system not too long ago that look for various criteria and be verified based on that. it is not a function of how many
3:28 am
followers, we have some verified folks who only have 5000 followers. >> that was his point. i don't know what to tell him. from what he explained to me and to staff is that they applied at the same time. why was he denied and they were approved? >> i would need to understand his particular case. i would want to know his name. >> we will get you the information, because i would like to get the young man an explanation. >> ok. ,> and let me ask you something and i apologize that sometimes you are a little redundant being the last one. committee has been the tip of the spear with the opioid crisis. aware, we are losing 115 people every day to opioid addiction. and we just talked about the algorithms, and you have been talking about it all day.
3:29 am
why is it that we have not been able to get these salt -- solved? what are you missing to not be able to get these tweets off? it is more of a new behavior and new approach. >> this has been going on it while. >> it is certainly not an excuse. we need to look at these more of how hererms algorithms are determining when we see this activity and take action faster. >> fair enough. my last question is this. i want to talk about intellectual property, particularly how it relates to livestream. the senate this afternoon, and all day long. you have been saying, and we have no reason other than to believe you, that we will work on this.
3:30 am
why should i believe you, and we had the ceo of another social media that was here a couple of months ago, same thing. we are working on it, yet this is something that has been going on and you haven't got it done yet. i echo the comments, i don't want the federal government to get into this business. i don't want to regularly -- regulate you guys. if why should i believe you you haven't gotten this fixed? the reason we have to still work on it is because the methods of attack constantly change and we will never arrive at one solution that fixes everything. we need to constantly iterate stealing iptors of
3:31 am
or rebroadcasting ip because they are constantly changing and we need to be 10 steps ahead of that. >> i want to believe you and i am going to believe you, but i just have to tell you, we don't want the federal government, and you don't want the federal government in this. i think the success of the internet and your products has been because the federal government stayed out of it, but we have to have help, we have to have commitment and when i look at this, why should i believe this. >> thank you, and while we have been sitting here, i am told twitter has deleted the account that was trying to sell drugs online. your team has been on work -- at work. you andausted probably your team and our members' questions for now. we do have questions for the record. want to thank you for
3:32 am
being here before the committee. we will be submitting those for the record and we have a number of things i would like to introduce, consumer technology association and the internet. article from nbc, article from slate, pursuant to committee roles, 10 days for additional response to the record. upon receipt, we ask that you remain seated until the twitter team is able to exit, so if you would all remain seated so the leave and twitter can thank you for being before the committee. with that, the committee is adjourned.
3:33 am
-- c-span, where history unfolds daily. in 1979, c-span was created as a public service by america's cable television companies and is brought to you today by your cable or satellite provider. we continue to bring you fromtered content
3:34 am
washington, d.c. and around the country. c-span is brought to you by your cable or satellite provider. c-span's "washington journal," live with news and policy issues that impact you. thursday morning, the wall street journal supreme court correspondent discusses brett kavanaugh's confirmation hearing reporter talks about foreign influence on social media platforms. the sure to watch c-span's washington journal. join the discussion. they want your attention, they want to feel important and they want to understand what is going on in the world. >> i worry is they become intimidated and feel they don't have they have those dinner
3:35 am
conversations, so they feel they can't throw their ideas out. >> hamilton is the rage. they love it. songs, they'll go on and on singing. let's have a contest and talk about the issues in the song. i think it will make it relevant . they enjoy it and they are engaged. >> saturday at 10:00 a.m. eastern, meet the civics teachers who participated in the c-span classroom annual educators conference. saturday at 10:00 a.m. eastern or on the free c-span radio app. sunday night on few and day, assistant editor at the atlantic zachary would talks about his and growing upd" in a troubled home. and iphone starts ringing see it is my mom. for some reason, i had a sense
3:36 am
is not going to be good. i don't know what it was, but i said this is not going to be good. i answered the phone. she spoke, she was very calm. zachary, child protective services are here. please come home as soon as you can and i knew that tone in her voice. child protective services is here. i asked myself if i was going to live to see the next day? that is what is going through my mind, because she, if ever at any point she has alone with me and i make it through whatever happens and i get home and talk to them, lord knows what she is going to do. on c-span's qd and a. >> early yesterday morning, british prime minister theresa may returned to parliament to
3:37 am
take questions from lawmakers for her weekly question time. it is her first appearance before the house of commons and the body recessed in july for their summer break. this runs about 45 minutes. >> order. questions to the prime minister. >> mr. speaker. thank you, mr. speaker. i am sure members from all sides of the house would like to join me in congratulating the english forscottish women's team qualifying in next year's world cup. missed youg, i have -- ministerial comments. i should have further such meetings today. mr. speaker, my constituents believe -- is


info Stream Only

Uploaded by TV Archive on