tv National Security Commission on Artificial Intelligence Conference - PART 1 CSPAN November 25, 2019 5:33pm-6:12pm EST
i am joined with two close friends of mine and i'm probably the only person who could say this in the entire world, i work with and for both of them but i want to disclose my conflict of interest to start with. general shanahan went to michigan go blue. rotc to enter the service of our country 1984. he has been promoted a gazillion times in charge of intelligence and operation activities and eventually we needed somebody off operational to have ai with the dod he was the perfect choice i worked with him as my role as chairman. kent walker was a federal prosecutor law and order federal prosecutor who chose to come to silicon valley working with ebay and google snagged him about 15 years ago.
we've worked every day together and during that time that only set up the legal function but in charge of all global policy. very significant players. what i thought we should do since you have heard from me is simply start and perhaps kent had to make some comments about the world you see today. >> thank you very much eric and general shanahan it's a pleasure to be here today with all of you. the partner on - - with this panel public-private partnerships is important i grew up in this area my father in the service 24 years for quite spent the first years of my life on us military bases my father had his career so i have commitment to get this right to make sure the private sector and the tech sector and universities can work together.
before we jump into thoughts of how to accomplish that i want to take on two issues up front it is frustrating to hear concerns around our commitment to national security and defense so i want to set the record straight. first china. 2010 remember google was public with the attack on our infrastructure that originated in china. with that cybersecurity attack while a number of our companies had significant commercial operations in china we chose to pick operations they are very carefully the focus is on advertising and open-source platforms. second more general of national security and our engagement in the project it is an area that we decided to wait until we had our own
develop our own ai and review process. that was a decision focused not a broader segment of the willingness to work with the department of defense and national security administration. to continue to that we are committed to doing that and that is on a long tradition of work on national security generally. so remember the history of the valley is on government technologies with radar and internet and gps to autonomous vehicles and personal assistance just now in the last couple of weeks we had an extraordinary accomplishment which move forward the frontiers of science and technology that was not an achievement by google alone it was built on research that had been done at the university of
california santa barbara to benefit consultations at nasa and carried out in many ways on supercomputers from the department of energy. those kinds of exchanges and collaborations are key to what made technological innovation as successful as it has been. and we feel weird contributing to national security a lot of that community is a part of google we go and above and beyond while having careers and even the tools to take steps to transition to civilian life to make use of their military skills in the private sector. as we do that we are fully engaged in a wide variety and working on a number of
national initiatives from cybersecurity to healthcare and with the darpa we are working on fundamental projects to ensure ai and to progress the operation of hardware and use software and hardware interfaces in a better way. as we take on those things we want to do more and pursuing actively those occasions to allow us to engage and we think that's extremely important for quite the same time a great partnership to be had the ai principles it is a lengthy document and continues to lay the groundwork by the department of defense back in
2012 with directive 2009 talking about the application of advanced technologies in the work dod has done. and the private sector also trying to drive forward on this to be very common in overlapping areas safety , human judgment, accountabilit judgment, accountability, explan ability, fairness those are all critical areas with different actors and each have different things to contribute and that's important this is a shared responsibility to get this right and we need a global framework to endorse that framework around these issues and something we want to support in working together because we are a proud american company committed to
the defense of the united states our allies so they can secure the world and we will continue the work and thinking of the places we can work together to know about everybody's strengths. >> thank you can't. general take us through. >> first of all it is great to be here thank you for the opportunity to do this i am a poor substitute for the chairman joint chief of staff although i don't have any headline grabbing soundbites. [laughter] also undoubtedly the first and last time i will serve as a warm-up act for doctor kissinge kissinger. [laughter] hang on for the main event. [laughter] i not only welcome but relish the opportunity to have a broader conversation of public private partnerships. reflecting over the director of project made, there is one
overarching theme and the importance and necessity to strengthen bonds between government and industry and academia. you brought up in others mentioned it but the idea of the relationship should be detected as a triangle in the form of the equilateral triangle government academia demand industry but i would suggest it did take that form in the fifties lasting until the early part of the decade walter isaacson writes about this eloquently and powerfully in his book. it is what really drove silicon valley today. but at best the sides of the triangle are no longer equal distance maybe they are
distorted or a little frayed to be different links the reasons are complex. and with the different business model with government and industry so the task is made much more difficult today that industry is so much more than department of defense with the adoption integration of ai. and those of the tech industry with the department of defense. and is far more than portrayed. we don't make it easy for them. so of those themes in the security commission's repor
report, that is the idea of a shared sense of responsibility upon the ai future of the importance of trust and transparency. so the national security depends on that. for various reasons viewed with suspicion are reluctant to have that strategic competition with china and ai is a critical component of nation's prosperity and vitality and sufficiency. so no matter where you stand with respect to the government with ai technology i submit we can never attain a vision outline without industry and academia with equal partnership. there's too much at stake to do otherwise. we are in this together of the
essence of the nation. but also across the united states government. we have to make this triangle back to what it used to be. >> thank you general i will ask a couple of questions to both of you. talk about similar. [laughter] >> and also an enterprise company and there are different protocols and ways of engaging. that all employees have the identical view but they don't.
it's a positive as well as a negative. in that constructive debate is america's first innovation. and quantum mechanics. and out of that comes incredible strength. and then to follow a more resilient framework so as he put for the ai principles in the governing process and that the report devotes a couple of pages of principles. so a lot of the hard problems conflict in our challenging.
but we have had debates whether or not to publish papers on lipreading that is a great benefit but that could be misused for surveillance and other purposes. that it is appropriate to publish because that particular technology was used only for one / one. 's love those discussions that you have and other challenging questions we have to come to terms with the reality and the trade-offs we are making very much the case and a lot of these issues as well. collaboration and coordination with cybersecurity and logistics and transportation and healthcare and those that are already engaged with the military.
>> when we started project leave and we went after commercial industry and eric said this is a solution that does already exist and do not reinvent the wheel. we wanted everybody in the market to the biggest internet data cybercompanies in the world why do we have to google project maven because he want the ai talent against the most wicked problems. those are extraordinary difficult problems to go after and we have a successful collaboration on this. so how that played out in a different story we got all the way to the end and that which we were very pleased with.
almost where you feel ostracized for working with the department of defense and those working on the project. and what we found is the critique on both sides as we lost the narrative very quickly. our approach of department of defense is not to become public that what the company wants us to talk about. in very general terms. with an intelligent surveillance remotely piloted aircraft. and what the project made and at what it was and was not. no pun intended.
controversial has been inserted permanently. i say is not controversial to anybody right now so full circle i'm not sure everybody fully appreciates our reasoning and that is a canary in the coal mine the fact that it happened and to get that out of the way with that reset and all the companies that we deal with at the department of defense it would happen to somebody else at some point but the idea of transparency for what each side is trying to achieve.
>> we don't wear hats anymore because with three has trying to figure out which one i'm wearing so the real problem inside the military with the soldiers and airmen and put them in front of the mind numbing operational tasks to watch the screens all day. it's a terrible waste of the human asset the military produces. so it's a huge opportunity to get them to work at a higher level position with that joint center for ai. and next dealing with ethics in the middle of the kerfuffle in google there was a good idea to have a formal ai ethics proposal that has a
remarkable public document. in that is definitive and maybe you could talk about that. similarly and i believe you are the customer for the proposal on the ethics. i assume both of you are in favor now they have copied their approach. does it really work does google stop doing things and how does that work cracks the same question for you, general.
but we cite the rules the military is required to operate under. >> as a general noted to have those set of principles without review processes it is the internal and external transparency. and to talk about surveillance we want to make sure and then it is deployed in the appropriate way. and then to license at for commercial use. and that's valuable and when the expectations are clear. and with that facial recognition.
and that with policy and technological safeguards. another example we have said it's nation technology to be very careful of ai in this area. and then to be deeper and more understanding. with a remarkable degree of convergence now internationally we see the european regulations over the next 100 days. and how we build acceptance for next-generation technology. >> in the dod lens.
with that area of convergence and then to drive a stake in the ground and if we don't disagree let's get the conversation going. i can tell you a certainty that china and russia with the public hearings and discussion of the use of artificial intelligence. so people may question what the department is doing in my beer doing it. just to make sure we took into account all the voices of artificial intelligence. you spent time and attention
and after those years in uniform with the ethical use of the department of defense with a history along the way of emerging technologies so there are differences of artificial intelligence similar to every other technology in the department here is substantive differences. that's pretty good framework as a way to look at this with artificial intelligence or technology our approach and our training is in place and how we bring it in from the prototype. so now this report is from the secretary of defense what do you think of the reports for
the best possible starting point so we have to come up with that implementation plan it is departmentwide implementation plan putting something together the chief information officer of how we implement for the department of defense is not an overnight task but now we have an outstanding starting point. >> that's a wonderful framing for where we are like to push so open ai has arbitrary rewriting that they became concerned only with certain models as an example they said
no they just thought it was our good judgmen judgment. famously very early said that we will avoid that with the dangers. where will the industry and up with self-restraint cracks is a common set of prints of one - - principles with respect to being careful cracks how will that play out quex's spin security see that with a partnership of artificial intelligence and exchange of information on the work that's being done in. more infrastructure and framework and the use of artificial intelligence and the safeguards and checks and balances. but when the common groundwork
cracks we are on the path to do that. the communications platform and television to the internet with regulatory infrastructure and how you use these tools it is an extraordinary technology and it is understandable but it's so notable because now you see the collusion. >>. >> and the notion of a new kind of warfare and the term that you used as operational warfare. so take us through so what is new and powerful about this technology in the military context cracks.
>> i go back to as we were formed. and now that's the team to get away from the research picture. away from the research picture. >> [laughter] to make your are going to kill me. when she tell us what hal gore rethink warfare is. >> redefining who 20 years counterterrorism insurgencies. we are going to be shocked by the chaos in the we'll have venus and the function of the future fighting.
maybe playing in micro sentences that time. how do you envision a pie happening. sp algorithmic news algorithm. as you described earlier, how fast can we get inside of 70s decision cycle. col. john boyd, air force col., the author of the decision act. it's how you get through the disciple of decision-making which is really never about the designer. it's about the orient. but in future fight we are looking at, this will be happening who fast, if we are trying to do this by human against machines, and the other side has machines and algorithms that we don't, it will be unacceptably high risk of losing that conflict. this is challenging because part of what you are getting at, in the future scenario, how are people going to be assured that our algorithms are going to work as an intended. they don't take on other things. but we will fall back on, is the
starting.who the principles gave us. this will be test validation. we have to do a lot more work on the front and by the time that we know it is being fielded. we are really going to be at a disadvantage if we say we are going to be in human against machine, it will be human and machine on one side, and human and machine on the other but the temples, the fleeting superiority, the decisions will be made that fast. it is like the algorithm is the algorithm. a student to be the key question, is what happens when the whole scenario is faster than human decision-making. because military, when there is a threat in general, peoples check with their son. there's a son of a rule of engagement preserves all of this human judgment around some number of minutes. not some number of nano seconds. how will the military and justice procedures to deal with
is no possible threat. the sooner it will be driven by innovation will happen at the lowest possible if it. what we have to do and places is to get people the policies and authorities in the framework to do with the need to do. the innovation, the people that will see i have a solution to this and i am going to write code. i'm going to develop an algorithm. if you give me the date and the tolls the payment, all of a sudden thanks, we can do that. who is the idea, more decentralized than a lot of people are comfortable with today. and that brings risk. we are talking about higher risk, but it is either that or risk losing the fight. who it's his idea of decentralized development and decentralized experimentation and decentralized innovation. and as described the innovation was described this morning, we have got to give them the push from above,. >> nairn inference.
>> we are staying already, sort of efforts to destabilize where this information campaigns and the like, we can work together to recognizes patterns across the wider battlefield if you will. will be a better who everybody. >> do you have a model who how the industry, one of the themes of our whole conference is the industry and the government need to work together broadly. and obviously we have seen in general here but i'm really referring to the government as a whole. but more than the dob that needs ai. do you have a model how the industry should work with the federal government and the state government and the deities and who forth. sooner i only talk about two important elements of the report. and it the first is this notion of trying to build the application and new technology who in the second is the need who a global framework. that helps with that process. the third is the more operational administrative question. how do we make it as easy as
possible who new companies to enter into these kinds of partnerships. who a lot of the innovation and the cutting edge research being done in silicon valley has brought being done by large companies. it is being done by small companies. the rich ecosystem of credit innovation. and is challenging even who a company google size to start to get more involved in that environment. it is doubly difficult who some of these small companies. who as you look at modernizing procurement from the military side and congress as well to make that as quick as nimble and flexible as possible. i've response to their needs. looking at increasing funding across the board because that's traditionally been a really fertile ground who a lot of these collaborative prices to move forward. and looking at the human resources exchanges, a lot of authorities out there which authorize by the sector people coming into the government. but in practice, and it is harder than you would say. who a lot of that hard work on
the ground, is important in making this adjustment. >> who both of you, because we're going to be making recommendations, plus or minus are there specific things that we can continue promote private partnerships. who example, the dod has a ux, in a number of groups that work very closely with intel, extraordinary countries has played to our industry and the technology and to me personally. who the sum of all of that, do you have a model that always get the same question, to a specific things that would be helpful that would decrease the friction and increased cohesion between small companies and large cavities in the federal government, pew kermit, the dod pretty. >> there's who much that it started to happen of last couple of years, and places like the dau, compiled from combat. all insurgencies to get things moving red.
>> these are each small teams of soccer people inside of the deity that have had an outside impact and taint changing the procedures and an important air force who example, in some of the chaotic summons things like this. spoon it's been with this all got started but we need to have a systemic change across the department of defense. u.s. about one of the ways to do that. all day one of the biggest ones is bringing in talent from the outside from academia from industries that are key scientists. prime the government, start up interest on, and a chief technical officer, 25 years in valley. he comes in within 24 hours takes a different view of what we are trying to do. we need sabbaticals and people coming in from academia who a year or two and going back out to the outside. people in education within the industry and corporate intelligence. that is all beginning to happen. we need to go to the next loophole to really start to understand what we are each talking about. we've going out to the valentine hampton suites, only get who
far. the peer to peer relationships and discussions that are going to be more important than anything else. >> i say we've seen examples of it. with darpa, we are priming the pump with a lot of important areas. whether that is training or in models and simulation that help format the recruiting in a number of different areas. another important component of this is the it modernization. the kernels, is critical but it comes embedded within a larger environment of sites where. it is often very times difficult because you have security clearances and appropriate certifications. but all elements of that piece. who that combination of successful individual experiments and trial runs to build that the mill area deep up your loophole. but also the systemic change to make it easier to have wider adoption of these technologies. >> is time who us to finish up. my objective was to put to bed,
this notion that somehow silicon valley would it work with the military. and we clearly seen examples small companies and large companies and we can can send her to forward and build this collective between the private and the public partnerships. can you sort of summarize sort of the key take away that he would offer us the key message in word, why are you here and why did you make a special trip just to make this. >> i want to be clear, and will see what i said at the beginning, we are a proud american company and we are committed to the cause of national defense who united states of america who our allies, and who peace and safety and security of the world. the approach as we deal in approaching of advanced technology, we want to be thoughtful and make sure that we have frameworks and transparency and understanding as we move forward. i say that is a mission that the military and u.s. government
share. we are looking forward to working more closely in the future. >> who general, you never like these things but you are sort of top, the fellow that we are going to sort of make this change happen across 3.2 million people and $660 billion, an enormous bureaucracy. how are you going to pull this off. >> one person at a time. [laughter] estimate combination of the top down. on the previous panel, as it was said, he must have the full support from the leadership. to show it is a priority from the department. that is critical and you have to have the bottom-up innovation of people working who blow. there's no question that some of them are represented in this room. they are and they know what the future needs will look like. at week meet in the middle and give them the resources and tools to succeed. the last thing i'll see, this is intimidating. it's a daunting task. there is no way around it.
it is a multigenerational problem is going to record wire a multigenerational solution were not going to wake up tomorrow realize that we got this all right. going to have some bits and start and some successes and drawbacks but just keep piling ahead and with resources, and the commitment of the department behind us. i know we will get there. >> thank you. it is worth staying, i work with kent who 15 years and with my google hat on i will tell you that it cannot be more proud of the impact that he had our society and skill in the reach of our corporation. i say you can see this today. in general, i don't say bob could've chosen a better person believe this. our partnership with you over the last three years, you really have moved the resources and gotten the her name her name, gotten the attention and delivered, and there was no one before you. you are that person. who thank you both very much and thank you all. [applause]