Hello, everybody. Thank you for coming. I am jennifer whitham, the dean of engineering at stanford university. And i am a cochair of the Stanford Emerging Technology review, which is sponsoring the event today i wanted to thank secretary raimondo for hosting this opportunity to talk about their review and to have the panel that you are all here probably excited to hear what they have to say. Im going to briefly tell you about the review itself. And then we will move ahead to the panel. We created the Stanford Emerging Technology review, because we recognized a need to educate policymakers, industry leaders, and the general public in key emerging technologies. You probably realize that todays emerging technologies and their impacts are very complex, and they are evolving at an unprecedented pace. And our goal is to enable better understanding of the emerging technologies for the people who make the laws and policies around the technologies and for the people who develop them, market them, and use them. The review is a great example of a collaboration that really couldnt happen anywhere other than at stanford. It brings together our topnotch engineers, who i represent, with our experts in policy and government from the Hoover Institution and elsewhere on campus. At stanford and in the engineering school, we have a long history of collaborating across the campus. The collaborations have been key to everything from the creation of silicon valley, to some Pretty Amazing discoveries in human health. But the review is an example of a new type of collaboration, a capital a collaboration between engineering and the Hoover Institution which we hope will have farreaching national impacts. Our students and faculty have been eager to contribute to this project, and i guess more generally, to understand the societal implications of their work. I have seen an increase in this desire in recent years. The review, i have to say, is exciting and fun for others as well, collaboration that we hope will have big national impact. It is truly a collaborative effort, and i have been delighted to be part of it. I am now pleased to introduce first, the moderator of todays conversation, maryam vogel. She is the ceo of equal ai and a chair at the National Ai Advisory committee. And joining maryam are gina raimondo, the u. S. Secretary of commerce and two of my stanford colleagues, condoleezza rice, forming u. S. Secretary of state and director of the Hoover Institution at stanford, and professor faith elite. She is so Quiet Capital professor in Computer Science and codirector of stanfords institute for artificial intelligence. I bring to you our panel. [applause] maryam thank you for joining us today. In real life, virtually, thank you to our panel for being here today, for this truly important discussion. Todays section on emerging Technology Review tour with the Hoover Institution enables us to have a conversation about what i would argue are the most important topics of the day. We have been fortunate here to assemble the titans of industry, international leadership, ai policy, the trifecta. I did not think it would be worth noting, but perhaps for some in the media, it does turn out when you are assembling these titans, they are all female. [laughter] we will let that data point go for now, and jump right in to the important conversation about the topic of the moment, ai. What is governments role in shaping ai policy, and the critical and challenging balance between harnessing its opportunities and safeguarding against its harm. Lets jump in. Lets start with setting the table with that very topic, the broad topic, that we started off here, in terms of governments role in ai. Given you each have a unique Vantage Point in understanding what that role should be. Secretary raimondo, since this is your mission day in and day out, can you share what is your northstar in understanding governments role in shaping ai and its development . Sec. Raimondo thank you, first of all. Miriam is fantastic and is the cochair of the National Ai Advisory board, which i put together. And she is just extraordinary. Either way, parenthetically, this may be the only highlevel ai even with all women. I think that is well done to you. [applause] because one of my worries is that they are not nearly enough women in the highest levels of nominees that develop ai. And that is a problem. You have such a powerful technology, you need to have all views developing it. Anyway, as it relates to the government, im sure many of you who are technologists or businesspeople or students think, when you think government and technology, you may wall your eyes. With good reason. I just want you to know, we are running towards ai, not away from it. Even though government may historically be slow to adopt technology, this is too important. The upside and the downside. Most highlevel, we are running towards it, working in collaboration with industry, with civil society, with academia. And my northstar is to fuel innovation safely. The United States leads the world in ai. The whole ai stack. The hardware, the design, emodels, the cloud. We want to stay that way. And we lead that because we out innovate the world. We have to make sure that whatever regulation or guardrails we put in place never sacrifices that. And faith, i am sure, well talk about the promise of ai. The downside is pretty scary. The adage in silicon valley, move fast and break things, cannot be used here. My job is to make sure ai is developed in a way that is safe, sick or, and people trust. Safe, secure, and people trust. We are about to make new announcements around aia safety institute. Setting standards, what is safe ai . When you get the results of your red team, what is adequate . What is safe . Standards dont slow down innovation. If you do it right, you set standards and promote innovation. Mostly i am very focused on the National Security concerns. These models getting into the hands of nonfaith actors or people who are our allies are very dangerous. I will leave it there. We are excited. We are moving forward. But we really have to put in these guardrails because it is good for everyone. Miriam thank you, secretary. We jumped into the conversation because these women need no introduction, but interesting to underscore the insights that come to us from the secretary of commerce, given her deep background as a leader and policy as its governor of rhode island. Having been a venture capitalist, navigating these with skills that come organically. Then speaking of unique Vantage Point, secretary rice, you have been inside and outside government. Perhaps not when you were in government of ai per se, but related questions of innovation technology, what do you see as governments role with Ai Development . Ms. Rice thanks very much. I really do hope this is the beginning of a conversation, let the end of it. I think we have so much that we need to talk about that we need to explore. I just have to say, secretary raimondo has been at the forefront of understanding the paradox you just put forward. Which is that this innovation, these technologies, the u. S. Lead in them, that is something to celebrate, and very much want to protect and move forward. On the other hand, we know government and institutions also have a responsibility to its citizens, to the common good, to make sure we are doing the best, and be wise about how this is going forward. I think because of your experience. You have been very good at helping us to understand those tradeoffs. I would say, two things about the governments role. One for my National Security perspective, one from my perspective as the former provost of stanford. Why does the United States have the lead that it has in so many technologies . If you look back to the time after world war ii, the United States government actually made a bet. A man named Vannevar Bush had this idea that universities could become the platform for the exploration of fundamental technologies. At the time, we called it basic research. It did not have to the government should not be picking winners and losers. It was really through the National Science foundation, or through the nih, it would encourage faculty, graduate students, postdocs, to dream big about what was possible. Funding was given to dreaming big about what was possible. Out of that came some of the biggest breakthroughs in universities on some of the most fundamental research. Because the government understood that these things may not be commercialize a bowl commercializeable ever, but they certainly were not immediately. That protection of that fundamental base became one of the hallmarks of the United States. I can tell you as provost of stanford, i got to oversee that research enterprise. And of course, we have a lot to come out of that. There is a story going before that in 1934, these two graduate students asked their dean of engineering to give them 500 to commercialize their product. That would be bill hewlett and david packard. In 2000, the university would help a couple of its graduate students, larry page and 30 brennan, who founded google. The ability of the government to fund this fundamental research, and enrich the economy and made us the leader, i think understanding that proper role of government is very important. For my National Security hat, is a former secretary, i would say this. We have an adversary this time around that is unlike any adversary we have had in the past. That is in this race for these fundamental technologies, for these transformative technologies, and that would be the peoples republic of china. I emphasized the adversarial part of it, because when you listen to xi jinping talk about the fusion of Civil Technology and military technology so that china can take its rightful place in the world, which by the way, means displacing the United States to take that rightful place, one has to be concerned about the protection also of our intellectual property, what is being developed in our labs. Finally, bringing two worlds together, you also dont want to wall off the United States. Because this is going to continue to happen in labs across the world. How can we keep our advantage, press our advantage, and do it in a world in which we dont want to make it worse for humanity . Miriam thank you, secretary. Finally, to round out this important conversation by the best way we could round it out, the practitioner. We are so fortunate to have your deep insight as someone who has built ai programs, who has developed innovations, visionary ai programs, who has been in academia through stanford, through founding and cohousing the stanford high center, understanding the balance of the government role, and i think it feels weekly but it might be monthly publishing papers on the governmental. Sure with us your insight . Thank you. And thank you for this incredible panel. I think you mentioned that this is a fairly rare panel. I do want to say that i have been talking to secretary raimondo and secretary rice about ai for years. These are visionaries in our government, and our academia, who have a deep conviction to advancing our countrys ai technology. And butter the whole social economics of our country. Ai, you know, i have been doing ai for the past almost three decades. It started with the timing, intellectual curiosity. These were seeds planted in american universities. I remember i was an undergrad, and resourced by our government funding, and by our incredible ecosystem. These seeds blossomed. There are thousands of flowers, but some of them become incredible public goods. They become civilizational driving forces of change, and ai is one of them. At this very moment, while i dont know about you, those of us in ai in the past 14 months, i feel like i have not slept. Its very exciting to see the breakthrough of technology, the public awakening moment. It is exciting to be part of the acceleration. But i also have reasons to stay awake. That is a little bit of anxiety and tension. What america has been known for, and we do the best, is the incredible ecosystem. Just like the examples secretary raimondo and secretary rice cited. Decades and decades ago, our government, our private sector, our Public Sector, our academia, have been in this ecosystem nurturing and fertilizing ideas, and translating them into scaling opportunities, and leading our country. Today, not a Single University in america has the resource to train a chatgpt model. That is already yesterdays model. The lack of resources on our campuses, in the hands of students and faculty, is a very strong limiting factor. Why is it a strong limiting factor . Just like conzy has said, and the Early Research is seized for incredible research. Todays ai is a Horizontal Technology that can empower scientific discovery. Many of you know or have heard of the fusion break through a year ago. But you may not know that a lot of that was due to the help of machine learning. Many of you are in works for health care, drug discovery, environment, climate solutions, manufacturing, agriculture. So many of these sectors are going to benefit from the power of ai, from the power of Ai Innovation that starts in academia and Public Sectors. To answer your question, i think the most Important Role that i would like to help bridge the conversation, is to encourage our country to adopt that mentality in Public Sector investment of ai. Starting with amazing progress that im sure secretary raimondo will talk about, like the National Ai Research program, and some and some of the bills that is being considered in congress and the senate. Miriam thank you so much. Now lets open our aperture a bit. We know when we are talking about ai, it is a technology that crosses borders and nations as it is being developed, deployed, and used. Fortunately, that was not lost on anyone on this panel or our government. As we have seen in the recent executive order on safe, secure, and trustworthy ai, numerous mandates within the executive order to Foster International ai safety. We have numerous efforts that build on our u. S. Trade technology council, where secretary raimondo cochairs with secretary blinken. And our representatives, arent we fortunate, as well as the g7 hiroshima project and several others. Secretary rice, given your former and current role in international relations, why, in your perspective, is International Collaboration a priority . First, i should start with, is it a priority . And if so, what would you like to see the government doing right now . Ms. Rice im going to subdivide. Because i no longer think international is the right word. Im going to start with trusted friends. You have mentioned the eu. We, of course, have friends in other parts of the world in asia and the like who share certain values. Because one of the things that is hard about this technology, ai, is that it will and body the values of those who use it. So we know that when you are dealing with ai in the hands of authoritarians, it is going to be used for different reasons, for different purposes, van and our distributed everybody will question, there will be congressional hearings. That does not happen in authoritarian societies. While i would like to look toward one day broad Scale International cooperation, i would really like to start with friends and allies in understanding what we even mean by standards and collaboration. I often say to my european friends, we talk about privacy, and we seem to fight about privacy all the time. Whatever differences we may have about privacy pale in comparison to the differences we would have with authoritarian regimes which basically dont have the concept of privacy. I will start in the trusted environment and then try to work out. One of the places i think you can outside of the trusted environment, i know you are working on standards. Standards get embedded in all kinds of international domains. One thing we dont pay enough attention to in the United States is standards that get written in various u. N. Hotties and various international bodies. When i was secretary, somebody would say, we need somebody to go to fill in the blank standards meeting at the u. N. And i would think, ok, what retired or nearly retired officer can i get to go do this . Because i cant have any of my key people go to it, i am too busy. That is really the wrong attitude. Because what has happened is a lot of these International Standards the chinese on the russians were not thinking in that way, and they were putting people into those positions. I think we have to think differently about that piece of the international environment. While we are working with our friends and allies who share our values of what might be harmful about ai. That is a bifurcation of the task. Miriam thank you for that strategic vision, and renaming. We always appreciate having that clarity, as well as thank you for the honesty and how your perspective has changed. Thats really helpful. Secretary raimondo, we mentioned how you are leading in the space. We only tension and a few of the ways you are leading in our international effort, including representing us at the u. K. Safety summit last november. You have written extensively speaking on this. You have an oped. You cowrote with secretary blinken last summer on how we need to act now to make sure we are appropriately shaping the future of ai. Why, in your mind, has International Collaboration been a priority and what do you hope to see happen during your tenure . Sec. Raimondo we run the National Institute of standards, at the commerce department. Science and technology, but it is all about standardsetting. It is called nist. We have got hundreds and hundreds of people who live and die by standards so they will appreciate that you have seen the light on the value of standards. We cannot i mean, the secretary is right. We cannot do this in a vacuum. Ai models, travel models, can travel International Borders very easily, open source models can be used by anywhere, anytime, all over the world. I agree completely with condoleezza. We have to have a group of countries who share our values, share our values as it relates to democracy, human rights, openness, protection of privacy, protection of basic rights. And together, come up with standards. What is safety . There is a lot to talk about red teaming. What are the results we are looking for. What are the dues and the donts . How can you use and develop ai to commit human rights abuses and the like . Here is the exciting thing. We are in the early innings it is interesting. Feifei said she has been at this for decades, and it is true. We first met 10 years ago when i was governor, and i reached out asking how we could use ai to better help older folks in rhode island Access Health care. We became to be friends. Ai using homecare. That being said, generative ai, and the models we have now, were in the early innings. There is an opportunity right now to align our rules, regulations, policies, and standards with europe, as you said. Japan, india, other democratic countries who will share our values and write the rules of the road together. We have already signed an mou with our ai safety institute, with the u. K. Safety institute. We have a trilateral discussion going with japan and south korea around ai and it standards. On the one hand, this is all daunting. On the other hand, we are kind of all going at it at the same time. And i have high hopes we will be able to align. Miriam thank you. Speaking of opportunities, and following on your point about mayor. You mentioned the national ai resource. You were one of the first in the world to call on the u. S. To create mayor. You served on the task force, talking about how to establish it and why. We have the eo that has an Important Development of the pilot of it. I am compelled to mention that the National Advisory committee included in our april report recommendation that the mayor be piloted as well as established and funded. It seems the rare legislation that has broad bipartisan support, little to no opposition. Can you explain or anyone who is not familiar with it, why do you support it . What would happen if it were to come into existence . Who would be the beneficiaries and how do we make that a reality . Dr. Li thank you for mentioning it. How did it come about . About five years ago, with the support of several leaders like condoleezza and jennifer, and our president s. We establish the human center ai institute, and our Founding Mission is to better ai, advance ai research, education and policy, to better human conditions. Michael director my codirector, we have very good provosts, we wanted to tackle a critical fundamental challenging problem in ai Public Sector academia that would answer to the core issue. We looked around, this was before jan ai. Gen ai. We recognized the lack of compute, which is chips, lack of data, is the fundamental problem that is suffocating Public Sector ai and academia. Because of that, in 2020, we wrote a public letter and we mobilized dozens and dozens of leading Research Universities around the country. And advocated for a bill. The nice thing about working with former provosts is every president of the university in the country are on their speed dial. And we got bipartisan support, we got private sector support, and we advocated for the congress to pass the mayor Task Force Bill in 2020. And President Biden called on the task force to work on the planning for the next 20 months. I was a member of the task force. And early 2023, we finished the planning and recommended at least 2. 6 billion effort in setting up the National Ai Research cloud, where we can provide the computing chips and computing hardware, as well as Data Repository to the academia, where we can develop the next generation of cuttingedge ai research. We can also focus on bringing the proper way of thinking about trustworthy ai, responsible ai, ai safety, into the development of ai. And the eo, fink you, miriam, for your wonderful work, thank you secretary raimondo, but the eo by President Biden included a Pilot Program of mayor. I am sitting here very anxious because on tuesday, the Pilot Program was open to researchers like me while all my colleagues are applying for it. Washington, d. C. , you know. I might be late applying for that mayor credit. But we are very excited. This is only the beginning. This is not enough. We really hope that we will finish the job by passing the create ai act which will appropriate the the proper funds for the full mayor resource. And enable and empower our academia and Public Sector, and truly come the american ecosystem that makes us the leading country of innovation. Miriam how much do you think it has to be for mayor to be what it is supposed to be . Dr. Li let me just give you an example. By 2019, one big tech company, namely microsoft, put in more than 10 billion in one startup, namely, openai, to train their gen ai model. This is back in 2019 and 2020. I personally think we are looking at billions and billions in the scale, tens of billions, in order to do this right. This is why condy and i talk a lot about this moonshot mentality. There is foundational infrastructure, there is data infrastructure, and there is possibly ideas like training talents, regional labs, national labs. I would love to see that full program. This is such a great panel. I wanted to double down on something you all have mentioned. Huge effort by secretary raimondo. Congress passed the chips act in 2022. The part no one talks about is that it authorized 174 billion more fundamental science. It has been authorized, but the appropriations are lagging every single. A single year. Now they are coming up with a new bill to do it for ai. It gives you the opportunity to Tell Congress you care. I will say one thing on the, and that is why i was asking you. Fair is fair, they appropriated some of the science money. And i do want to say a huge announcement this week, we have announced a woman, who is like the number three person at synopsis, is going to be running the National Semiconductor technology center. That will be funded with billions of dollars through the chips act to do exactly what you guys are talking about. Fundamental r d, fundamental r d, and more workforce training with universities and companies, focused on chips. Credit where credit is due. I am but what everyone has said here, you cant do the moonshot without the fundamental research. We ought to be able to agree on a bipartisan basis that investing in research and development, to enable america to lead the world in fundamental technology, that is just good for our country. It is good for our National Security. I will let condoleezza speak to that. We cant lead the world unless we are the most competitive country on the planet. I rise and sleep every day thinking about competitiveness, how do we compete. Its innovation, workforce, protection of ip, it is a whole bunch of things. Fundamental research, it is kind of embarrassing that what fei fei said, there is not a Single University in our country that has the resources to do this modeling. Fund it. In brief, to congress, fund it. It is bipartisan and it will help us thrive. And i would add, there is a cautionary tale from this chips which is the United States wants once dominated that space. We created it, dominated it, and we looked up one day, and we lost it. We lost to leadership. We also looked up one day and the supply chains were not what we thought were. We thought they were. If you are not paying attention, you could lose the lead in these areas just like, particularly when you have determined competitors in the prc, for instance, who have whole government ways of dealing with this. Dr. Li mai space, so many countries have ai strategy in the late 20 teens before the United States had a national ai strategy. That is an example. Miriam i know you talked about how there is so much knowledge, not enough wisdom. On this panel, that proves that. I hope people are taking notes, and that we will be able to followup on all of these important points. Back to this executive order, secretary ramonda, it seems almost half of it is directed under your purview, which is significant, since thanks to a report by stanford hi, keeping track for all of us, there is over 150 action items of science. More than 50 Government Agencies and related entities. Secretary, you also mentioned balancing opportunity, balancing the risk mitigation, and making sure that this technology is safe for all of us. Can you share with us as you approach the enactment, it is seminal today, it is the 90 day mark. Most executive orders, it would be another day. In this executive order, we have a bunch of deadlines coming due this week. How are you approaching with some of these mandates that balance . Sec. Raimondo at the congress department, we are focused on both sides, paying offense and defense. Offense is investments. We are making huge investments in ships and semiconductors, etc. But the defense when it comes to ai is just as important, to keep us all safe. We are about to launch an ai safety institute, soon be making some announcements around that. In nist, just to do exactly what i said, set the standards with our likeminded countries around what safety is. We have already received voluntary commitments from the Biggest Ai Companies around how they will develop their models, what data they will share with us, etc. But today actually, coincidence and lucky that i am here to break this news, we are doing two things. One is we are using the defense production act, which is the authority we have to require to do a survey to require companies to share with us every time they train a new large language model and share with us the results. The safety data. So that we can review it. That is invoking the dpa. The second thing is we are beginning the process of requiring u. S. Cloud companies to tell us every time a nonus entity uses their cloud to train a large language model. These are two specific things that we are announcing now. It is going to be a pretty rapid cadence of announcement, coming out of the congress department, focused on safety. We need to know this. We need to know, what are you training, when are you training, what are the results, whats the data, what foreign actors, malign or otherwise, are trying to get access to your cloud to train their model. This is what we are working on now. If we allow bad things to happen with ai, nothing will stifle innovation more than that. Some people say, you are over regulating. How is it good if we allow ai to run amok, get into hands of nonstate actors, then the hammer will come down and really stifle innovation. Like i say, we are trying to walk that balance of speeding up innovation, but doing it safely. Miriam thank you for doing that. Im so sorry, this will be our last question. I would love to have this conversation all day. You all have talked about the values here. We are talking about technology that incorporates the values of those building it, of those adopting it, of this deploying. And you have all spoken, and a lot of your life work is making sure that american values, equality, opportunity, individual freedoms, are built in. That we are mindful as we deploy and develop ai. I guess lightning round, given the time, and each of you please share, what is the most important thing you think government could be doing now to make sure we instill those values in our ai . Is there something those of us in private government that can do to support . Sec. Raimondo i will start and i would say i would reverse it. There is not so much government can do to instill value. Our constitution, our system of government actually, less the values protection with the people. Even more with states than they did with the federal government. Because actually, our founders were kind of suspicious of whether governments would ever protect values. If you read, they mostly wanted to make sure the government did not get into that business. What can we do not in the government to do it . I would put it that way. We could start with education. Our students at stanford, and im sure it is true of students at m. I. T. Were george washington, or georgetown. They were born into this world. They are at one with this world. But we have to make sure that they are educated and trained in a way that they also think about the values that are both potentially advanced by these technologies, and potentially abridged by these technologies. I will let feifei say this, but we talk a lot about any class in Computer Science, or any class in ai, ought to be raising with our students questions of bias. It ought to be raising with our students what happens if this falls into the hands of authoritarians rather into the hands of democracies . Rather into the hands of democracy. This is the ultimate Civic Education problem. How to think about technology in a democratic society. I start with education, i think it is great jennifer mentions this, we get such great uptake to be a part of this from our scientists. I think they are concerned about what the impact will be on society of the technology they are developing. I will go with society to try and make sure we are protecting those values. We have our marching orders. I will say, george schulz, who died at 100, said democracy is not a spectator sport. On this one, democracy is not a spectator sport. I want to chime in on that, i will be brief. Im issuing a call to action. The work we are doing at commerce, we need the private sector, we need companies to help us, or resource of companies, talon of companies, talent from universities. The talent from startup. The compute power. Government we need it funded, for sure. That is what government can do. But really this is a all hands on deck moment. I cannot agree more with secretary rice. I tell all of my students theres nothing artificial about the field of artificial intelligence. Machines do not have independent values or human values. And america is a country that leads in our value. Not because the government told us, it is because the citizens of america hold those values together collectively. It is all hands on deck for innovating this technology. All hands on deck for designing our values into technology, and for us as educators, the nextgeneration ai business leaders, government leaders. Not necessarily english and spanish or anything else, but bilingual in terms of technology and social policy, humanistic understanding. Im so sorry for this conversation to be over. Weve had pearls of wisdom, deep insight from those at the frontline. Weve had a breaking news and call to action. Thank you all for squeezing that into the short time we have together. [applause] [indiscernible]