Sen. Chuck Schumer Launches SAFE Innovation in the AI Age at CSIS

This transcript is from a CSIS event hosted on June 21, 2023. Please cross check quotes against the event recording here.

Thomas J. Pritzker: Can you hear me? Yeah, this works? OK.

So welcome to CSIS. Thank you for coming. I’m Tom Pritzker. I’m chair of CSIS. We are a bipartisan national security think tank, and that’s why Senator Schumer chose CSIS for this speech. It’s an honor and pleasure to introduce Senator Schumer. I don’t need to go through the senator’s résumé, but I would like to go through some quick background.

My background with Senator Schumer is that he is both a thinker and a doer, and that makes him a bit of a rare person in most rooms. And for this moment of artificial intelligence, those characteristics may well be mission critical. In order to be a modern economy, we must have innovation driving both the creation and the adoption of technology. Our economy is going to depend on it. The economic – our economic growth and our national security will depend on the dynamism of innovation and of the private sector.

You know what Senator Schumer did with the CHIPS Act. You may not know what he did for quantum computing, so let me tell a quick story. The pre – in pre-pandemic time, we had a small private dinner here at CSIS with a group from the University of Chicago and from Argonne. They were building out an effort around quantum computing. Senator Schumer attended the dinner and listened carefully. And he listened not only to what they had to say about their need, but he also asked: Why Chicago? Senator, that seed has grown. DOE subsequently created the National Quantum Initiative, and that team became one of the five Centers of Excellence in the U.S. You know that part.

What you may not know is at the most recent G-7, IBM and Google announced $150 million of grants for a partnership between that team and Tokyo University to advance that same quantum computing. We now see an example of the government as a catalyst for the – for private capital to partner with the university world to further innovation.

I give this as an example of a promising way forward for American innovation. There’s enormous energy in our universities and in our high-tech industries. The government should encourage this unique American strength in a much wider and more systematic way.

The foundation of our real national security is the dynamism of our economy, the creativity of our research institutes, the quality of our education, and, importantly, the agility and skills of our private sector. If we don’t nourish these with a growth mindset, we’re playing defense.

Today we’re here to talk about artificial intelligence. Over the past two months I’ve been at two conferences that focused on the topic. The first was a conference organized by Cal Tech and University of Chicago around the use of AI for science. The participants were from major research universities and others around the country. In that room, the participants were palpably excited about the incredible benefits to society that would come from artificial intelligence.

More recently, I was at a conference of venture types around the topic of AI. Those are the guys who hope to make money off of this. In that room, interesting, the participants were palpably concerned about the future of what we were going into.

The problem is they’re both right. Navigating this is one of the great challenges and opportunities facing the United States. We’re at a pivotal moment. You’ve seen the CHIPS Act. You’ve now heard about the senator’s engagement with quantum computing. What the senator is here to talk about today is his approach to artificial intelligence, which may be one of the greatest challenges we face in decades to come.

So Senator, thank you very much for doing this. (Applause.)

Senator Charles Schumer (D-NY): Well, thank you, Tom, for that kind introduction. And thank you, CSIS. It’s great to be here. It’s a pleasure to be in a room full of so many leaders from the world of innovation and tech – business, academia, labor, civil rights, the arts.

Now, friends, we come together at a moment of revolution, not one of weapons or of political power, but a revolution in science and understanding that will change humanity. It’s been said that what the locomotive and electricity did for human muscle a century and a half ago, artificial intelligence is doing for human knowledge today as we speak. But the effect of AI is far more profound and will certainly occur over a much shorter period of time.

The idea of AI is not new. Supercomputers with humanlike behavior have long been with us, in the movies, in science fiction, in art. But now what once lived only in our imaginations exists in our day-to-day lives.

Thanks to remarkable innovations in computing power, in the speed of our semiconductors, in the size of our data sets, in the fields like machine learning and neural networks, we can say with confidence that the age of AI is here, and it is here to stay.

And we’re still just at the beginning. Some experts predict that in just a few years the world could be wholly unrecognizable from the one we live in today. That’s what AI is – world-altering. Change at such blistering speed may seem frightening to some. But if applied correctly, AI promises to transform life on earth for the better. It will shape how we fight disease, how we tackle hunger, manage our lives, enrich our minds, and ensure peace.

But there are real dangers too – job displacement, misinformation, a new age of weaponry, the risk of being unable to manage this technology altogether. We have no choice – no choice – but to acknowledge that AI’s changes are coming and in many cases are already here. We ignore them at our own peril.

Many want to ignore AI because it’s so complex. But with AI we cannot be ostriches sticking our heads in the sand. The question is what role does Congress and the federal government have in this new revolution. Are we capable of playing a proactive role in promoting AI’s growth? Can Congress work to maximize AI’s benefits while protecting the American people and all of humanity from its novel risks?

I think the answer to these questions is an emphatic yes. It must be, because if government doesn’t step in who will fill its place? Individuals in the private sector can’t do the work of protecting our country.

Even if many developers have good intentions there will always be rogue actors, unscrupulous companies, foreign adversaries, that will seek to harm us. Companies may not be willing to insert guardrails on their own, certainly not if their competitors won’t be forced to do so.

That is why we’re here today. I believe Congress must join the AI revolution and we need your help. AI is unlike anything Congress has dealt with before. It’s not like labor or health care or defense where Congress has had a long history we can work off of. Experts aren’t even sure which questions policymakers should be asking.

In many ways, we’re starting from scratch. But I believe Congress is up to the challenge. The last two years in Congress have been the most successful of the last 30 years: historic infrastructure legislation, the CHIPS and Science bill which Tom referred to, the American Rescue Plan, and the largest clean energy package ever.

Most of these were done with bipartisan support under my leadership as majority leader. So don’t count Congress out. I know many of you have spent months calling on us to act. I hear you. I hear you loud and clear, and many of my colleagues from both sides of the aisle hear you loud and clear as well.

So, today, I want to outline a two-part proposal to move us forward on AI, one part on framework, one part on process.

First, Congress needs a framework for action. What should our framework be? What issues within AI should we look at to prepare legislation after months of talk with over a hundred AI developers, executives, scientists, researchers, workforce experts, advocates?

This morning I’d like to share my proposed framework for action. I call it the Safe Innovation Framework for AI Policy. The Safe Innovation Framework. I call it that because innovation must be our North Star. I call it that because the U.S. has always been a leader in innovating on the greatest technologies that shape the modern world.

But if people think AI innovation is not done safely, if there are not adequate guardrails in place, it will stifle or even halt innovation altogether. So it is safe innovation we must seek.

Second, Congress will also need to invent a new process to develop the right policies to implement our framework. AI moves so quickly and changes at near exponential speed, and there’s such little legislative history on this issue, so a new process is called for. The traditional approach of committee hearings play an essential role, but won’t on their own suffice.

We will need help from creators, innovators, and experts in the field. That is why later this year, I will invite top AI experts to come to Congress and convene a series of first ever AI insight forums for a new and unique approach to developing AI legislation.

I’ll talk a little bit about these forums in a moment, but let’s return to the framework first. Let me repeat: Our framework must never lose sight of what must be our North Star, innovation.

America’s by nature a country of innovators. We produced over 590,000 patent applications in 2021. And 60 percent of the top 100 companies by market cap are American. It was America that revolutionized the automobile. We were the first to split the atom, to land on the moon, to unleash the internet, and create the microchip that made AI possible. AI could be our most spectacular innovation yet, a force that could ignite a new era of technological advancement, scientific discovery, and industrial might. So we must come up with a plan that encourages, not stifles, innovation in this new world of AI. And that means asking some very important questions.

One, what is the proper balance between collaboration and competition among the entities developing AI? Two, how much federal intervention on the tax side and on the spending side must there be? Is federal intervention to encourage innovation necessary at all? Or should we just let the private sector develop on its own? Three, what is the proper balance between private AI systems and open AI systems? And finally, how do we ensure innovation and competition is open to everyone, not just the few, big, powerful companies? The government must play a role in ensuring open, free, and fair competition.

In short, the first issue we must tackle is encouraging, not stifling, innovation. But if people don’t think innovation can be done safely, that will slow AI’s development and prevent us from moving forward. So my SAFE Innovation Framework calls for security, accountability, protecting our foundations, and, lastly, explainability – the last being one of the most important and most difficult technical issues in all of AI.

First comes security – for our country, for American leadership, and for our workforce. We do not know what artificial intelligence will be capable of two years from now, 50 years from now, 100 years from now, in the hands of foreign adversaries, especially autocracies, or domestic rebel groups interested in extortionist financial gain or political upheaval. The dangers of AI could be extreme. We need to do everything we can to instill guardrails that make sure these groups cannot use our advances in AI for illicit and bad purpose. But we also need security for America’s workforce, because AI, particularly generative AI, is already disrupting the ways tens of millions of people make a living.

At greatest risk are those who live paycheck to paycheck, displacing millions of low-income workers, many from communities of color. Trucking, manufacturing, energy production could be next. And, rest assured, those with college educations and advanced degrees will be impacted too. AI will shape and reshape the knowledge economy, impacting workers in sales and marketing, coding, software development, banking, law, other skilled professions. Many assumed these jobs would always be safe, but that is not the case.

The erosion of the middle class, already one of America’s most serious problems, could get much worse with AI if we ignore it and don’t take measures to prevent job less or maldistribution of income. Globalization is a good cautionary tale. Many heralded it as a turning point for prosperity and growth. Decades later, most people agree, globalization on balance increased wealth, but at the cost of tens of millions of jobs shipped overseas. While some communities flourished, others were hollowed out and remain so to this day.

Congress was far too slow to aid Americans who needed help with these changes. Let’s not repeat the same mistakes when it comes to AI. To prevent that from happening, we will need everyone at the table – workers, businesses, educators, researchers. This is going to be a huge challenge and all of us must be part of the solution. AI policies must also promote accountability. Otherwise, what will stop companies from using AI to track our kids’ movements, inundate them with harmful advertisements, damage their self-image or their mental health? What’s to stop a shady business from using AI to exploit people with addictions or financial problems, or serious mental illnesses? How do we make sure AI isn’t used to exploit workers or encourage racial bias in hiring?

And how can we protect the IP, the intellectual property, of our innovators, our content creators, our musicians, our writers, our artists? Their ideas are their livelihoods. When someone uses another person or another company’s IP, we need accountability to ensure that they get their due credit and compensation. Without guardrails in place regulating how AI is developed, audited, deployed, and without making clear that certain practices should be out of bounds, we risk living in a total free-for-all, which nobody wants.

Nor do we want a future where AI eats away at America’s foundations on its own. AI neither supports nor opposes the causes of liberty, civil rights, justice. If we don’t program these algorithms to align with our values, they could be used to undermine our democratic foundations, especially our electoral processes. If we don’t set the norms for AI’s proper uses, others will. The Chinese Communist Party, which as little regard for the norms of democratic governance, could leap ahead of us and set the rules of the game for AI. Democracy could enter an era of steep decline.

And there’s a more immediate problem. AI could be used to jaundice and even totally discredit our elections as early as next year. We could soon live in a world where political campaigns regularly deploy fabricated, yet totally believable and convincing, images and footage of Democratic or Republican candidates, distorting their statements, greatly harming their election chances. Chatbots can now be deployed at a massive scale to target millions of individual voters for political persuasion. And once the damaging information is sent to 100 million homes, it’s hard to put the genie back in the bottle.

What if foreign adversaries embrace this technology to interfere in our elections? This is not about imposing one viewpoint, but it’s about ensuring people can engage in democracy without outside interference. This is one of the reasons we must move quickly. We should develop the guardrails that algin with democracy and encourage the nations of the world to use them. Without taking steps to make sure AI preserves our country’s foundations, we risk the survival of our democracy.

Finally, explainability, one of the thorniest and most technically complicated issues we face, but perhaps the most important of all. Explainability is about transparency. When you ask an AI system a question and it gives you an answer, perhaps the answer you weren’t expecting, you want to know where the answer came from. You should be able to ask: Why did AI choose this answer over some other answer that also could have been a possibility? And it should be done in a simple way, so all users can understand how these systems come up with answers.

Congress should make this a top priority, and companies and experts must take the lead in helping us solve this very important problem. Because without explainability, we may not be able to move forward. If the user of an AI system cannot determine the source of the sentence, or the paragraph, or the idea, and can’t give some explanation of why it was chosen over other possibilities, then we may not be able to accomplish our other goals of accountability, security, foundation. Explainability is, thus, perhaps the greatest challenge we face on AI.

Even the experts don’t always know why these algorithms produce the answers they do. It’s often a black box. No everyday user of AI will understand the complicated and ever-evolving algorithms that determine what AI systems produce in response to a question or a task.

And, of course, these algorithms represent the highest level of intellectual property for AI developers. Forcing companies to reveal their IP would be harmful. It would stifle innovation, empower our adversaries to use them for ill.

But fortunately, the average person does not need to know the inner workings of these algorithms. But we do, we do need to require companies to develop a system where, in simple and understandable terms, users understand why the system produced a particular answer and where that answer came from.

This is very complicated but very important work. And here we will need the ingenuity of the experts and companies to come up with a fair solution that Congress can use to break open AI’s black box. Innovation first, but with security, accountability, foundations and explainability. These are the principles that I believe will ensure that AI innovation is safe and responsible and has the appropriate guardrails. If we proceed with these priorities in mind, I think Congress can help ensure that AI works for humanity’s good. I think we can go a long way towards keeping people safe.

Now let me share my second proposal – a new legislative approach for translating this framework into legislative action. Later this fall, I will convene the top minds in artificial intelligence here in Congress for a series of AI insight forums to lay down a new foundation for AI policy.

We need the best of the best sitting at the table – the top AI developers, executives, scientists, advocates, community leaders, workers, national-security experts – all together in one room, doing years of work in a matter of months. The panel will include people of differing views, including some skeptics. We want the experts in each subject where we have questions and problems to sit around a table, debate the major challenges, and forge consensus about the way to go. Opposing views will be welcome, even encouraged, because this issue is so new that all ideas must get a chance to be at the table. Our job as legislators will be to listen to the experts and learn as much as we can so we can translate these ideas into legislative action.

Each insight forum will focus on the biggest issues in AI, including, first, asking the right questions. AI innovation, copyright and IP, use cases and risk management, workforce, national security, guarding against doom-day scenarios, AI’s role in our social world, transparency, explainability and alignment, and privacy and liability. These insight forums are the first of their kind. They have to be the first of their kind, because AI moves so quickly, will change our world so dramatically, is deeper in its complexity, and lacks the legislative history in Congress that other issues have.

If we take the typical path, holding congressional hearings with opening statements and each member asking questions five minutes at a time on different issues, we simply won’t be able to come up with the right policies. By the time we act, AI will have evolved into something new. This will not do. A new approach is required.

Now, these AI forums can’t and won’t replace the activity already happening in Congress on AI. Our committees must continue to be the key drivers of Congress’s AI policy response, continue to hold hearings and build bipartisan consensus. But hearings won’t be enough. We need an all-of-the-above approach because that’s what AI’s complexities and speed demands. And this must be done on a bipartisan basis. AI is one issue that must lie outside the typical partisan fights of Congress. The changes AI will bring will not discriminate between left or right or center. It will come for all of us, and thus demands attention from all of us.

To deepen the spirit of bipartisanship, I’ve established a group of senators to lead on this issue – Senators Heinrich, Young, Rounds and myself. I thank my colleagues for their work. And I reiterate, while this is my framework and my vision for congressional action, I hope the same spirit of collaboration that we’ve seen so far will propel us forward in the months ahead. Of course we’ll rely on our committee chairs, once they hear from our forums, to develop the right proposals – Chairmen Cantwell, Peters, Klobuchar, Warner, Durbin, and their ranking Republican members.

Last week I asked each of the committee chairs to reach out across the aisle and together have the chair and ranking member identify and explore areas where we can get working. We also need the input of those senators who have spoken out on AI to join us, Senators Bennet and Thune, Blumenthal, Blackburn, Hawley, others.

Now, no question about it, this is all exceedingly ambitious. We must exercise humility, humility, as we proceed. We’re going to work very hard to come up with comprehensive legislation. Because this is so important, we’re going to do everything we can to succeed. But success is not guaranteed. AI is unlike anything we’ve dealt with before, and it may be exceedingly difficult for legislation to tackle every single issue. Again, humility is the key word.

But even if we can find some solutions and create a degree of consensus to deal with some of AI’s many challenges, we must, we must pursue it. And like many great undertakings in our nation’s history, we must move ahead with bipartisanship and cooperation. We must cast aside ideological hangups and political self-interest. That’s the only way our efforts will succeed.

In 1963, President Kennedy said in Frankfurt, Germany that, quote, time and the world do not stand still. Change is the law of life. And those who look only to the past or the present are certain to miss the future. What was true when John F. Kennedy spoke 60 years ago is even truer today. Change is the law of life, more so now than ever. Because of AI, change is happening to our world as we speak in ways both wondrous and startling.

There are many who think we’re in over our heads. There are those who fear AI’s immense power and conclude it’s better to turn back, to go no further down this unknown road. We all know it’s not that simple and it’s not that easy. The AI revolution is going to happen with us or without us. If we can promote innovation, make sure it is safe, if America leads the way, the future will be far better, brighter, and safer than if it happens without us. I do not know of any other instance in human history when we reached new heights, uncovered new truths or mastered new innovations only for us to turn back. It’s in our nature to press ahead. We are, as President Theodore Roosevelt said, the ones in the arena.

So friends, let us not turn back. Let us not look away. Instead let us forge ahead, determined, unafraid, to lay a foundation for the next era of human advancement.

Thank you, everybody, and I look forward to working with you on this effort very, very soon. (Applause.)

Gregory C. Allen: Well, Senator, thank you so much for being here at CSIS today, and for sharing the new SAFE Innovation Framework. This is a genuine landmark in the history of American politics, and AI technology, in the economy, and geopolitics. This really does touch on everything. And we’re fortunate that you were willing to spend the time with us today explaining it.

Sen. Schumer: Well, and I thank you for CSIS’s sponsorship of this.

Mr. Allen: Thank you. So we have a brief amount of time to ask you some key questions about this initiative. I’d be very curious to understand, you know, what was your process like? How did you develop this SAFE Innovation Framework?

Sen. Schumer: That’s a great question. And, first, I think we realized the breadth, depth, almost enormity of the task ahead of us. So instead of just plunging in with any preconceived ideas, we thought the first thing we ought to do is just talk to a wide range of experts in the field and in adjacent fields that are affected by AI. And so we spoke to probably more than 100 different groups of people. And we got their ideas. We got what their thinking was. We asked them questions about the framework.

And after we spoke to all of them, we felt confident enough – not with all the answers; that’s part of the forums that we’re going to do – but that we’re asking the right questions. And we feel quite confident that the balance we have chosen, North Star innovation but safe innovation so that the public just don’t pull back and say let’s forget about all this it’s too dangerous, that doesn’t happen. So we had a – you know, we spoke to so many different people, but almost everyone we talked to, when we told them that our framework is safe innovation, really thought that was a good idea.

Mr. Allen: And when you actually started the process of, you know, putting pen to paper, who did you work with to actually develop this framework, in addition to the listening with hundreds of groups?

Sen. Schumer: Well, that’s who we worked with. And I have a great staff. Some of them are here. And they spent so much time. And the enthusiasm on our staff – humility, but also enthusiasm of getting this done, and how hard it is to get it done but how important it is to get it done, has carried us. We’ve probably – I’ve had discussions about AI with my staff probably every single day. And they – knowing me, you know that includes Saturdays and Sundays.

Mr. Allen: So in this same room at CSIS about a year ago, we had the privilege of hosting Dragos Tudorache, a member of the European Parliament who’s been the co-author of the EU AI Act, which is advancing. In addition to the EU AI Act, we have very interesting proposals going on in the United Kingdom, in countries like Singapore, China, Brazil. What do you think of these other international efforts? And how will they, how have they influenced your work?

Sen. Schumer: Well, we’ve looked at them. And we certainly are going to reach out to other countries as we go through this process. But I think that none of them have really captured the imagination of the world and people say, oh, that’s it. Some of them are negative. Some of them have been withdrawn or modified already. And most of them were quite quick. So our goal is to come up with an American proposal dealing within the confines, if you will, of America, the United States.

And we believe, if it’s good enough, the rest of the world will follow. After all, we’re the largest economy in the world, we’re the innovative leader in the world, we’re the intellectual leader in the world. And most—not all, but most of the world would like to see one system. If we can put this together in a very serious way, I think the rest of the world will follow and we can set the direction of how we ought to go in AI. Because I don’t think any of the existing proposals have captured that imagination.

Mr. Allen: So the framework that you shared here today, the initiatives that you’re launching today, how do they compare with, you know, AI regulation in more closed societies, like China for example?

Sen. Schumer: Well, you know, it makes our task harder, but it will be more successful when we complete it. And that is, we have to include everybody. You know, I mean, I suppose in China, President Xi and a few others will decide what’s needed, and if they say there should be a lot of facial recognition, it’s done. That’s not going to happen here, just by who we’ve consulted, by the way we’ve set up these forums, inside forums. It’s going to be a very, very – a much broader group, and we’re going to have to pull a consensus together. Will that be easy? No. Will there be some areas maybe where there is no consensus? Yes. But that shouldn’t stop us on pursuing many places where we think we can find some consensus.

Mr. Allen: So one of the areas that you touched on in your speech was the importance of competition in the market and elsewhere. So how are you thinking about this issue of competition within the AI industry, and ensuring that this technology is not just dominated by a small handful of companies?

Sen. Schumer: It’s balance. On the one hand, you want innovation. And, you know, if certain companies be they large or small have that innovation, you don’t want to stifle it. But on the other hand, you want to ensure that many others get a chance at it. That’s one of the great questions we face. Some of it relates to OpenAI. But even without OpenAI, we have to decide those sort of guidelines. And I’d say it’s balance. But we certainly don’t want a situation where two or three companies dominate and everyone else is closed out.

Mr. Allen: And then, you know, we have a lot of people asking what is the timeline for this legislation that you’ve talked about.

Sen. Schumer: Oh, yes.

Mr. Allen: And how they can be involved in moving this process forward.

Sen. Schumer: OK. First question, the timeline. We’re not going to set an artificial deadline, it has to be done by this date. But I would give you this – I would say this: It’s not going to be days or weeks, but it’s not going to be years. Months would be the proper timeline that I would give you.

Mr. Allen: Great. Well, every minute that we have the opportunity to spend with Senator Schumer is truly precious, and his time is truly precious. So this is going to reach the conclusion of the event. But on behalf of CSIS, on behalf of the Wadhwani Center for AI and Advanced Technologies here at CSIS, thank you so much for spending the afternoon with us.

Sen. Schumer: Thank you very much. Thank you, CSIS. Thanks for coming, everybody. (Applause.)

Mr. Allen: I do have one final –

Sen. Schumer: I didn’t answer – I didn’t answer your final question. We need your help. How can people participate? We intend to be an open process. We intend through these forums to have different groups. We’ll have an opening one in September with some of the very leaders of the industry to start talking about the big questions that has to be answered. But then – I mentioned nine, there may be others. And we’re going to want broad participation and input, because many of you have expertise that we very much need.

Mr. Allen: Great. And so if everyone could please remain seated while the senator exits, to make it to his next meeting. Thank you, again.

Sen. Schumer: Thank you. Thanks, Greg, appreciate it. (Applause.)

Mr. Allen: Thank you.