Name: Brainstorm AI 2024: How Much Should We Regulate AI
Uploaded: 2024-12-10
Duration: 19 min 31 s
Description: Martin Casado, General Partner, Andreessen Horowitz, Dawn Song, Professor, Computer Science; Co-director, The Berkeley Center for Responsible, Decentralized Intelligence, University of California, Berkeley

Martin Casado, General Partner, Andreessen Horowitz, Dawn Song, Professor, Computer Science; Co-director, The Berkeley Center for Responsible, Decentralized Intelligence, University of California, Berkeley

Brainstorm AI 2024: How Much Should We Regulate AI

Conferences

Martin Casado, General Partner, Andreessen Horowitz, Dawn Song, Professor, Computer Science

Hi, folks. Thanks for being with us. Thanks for having us. Yeah, we figured we'd end with a nice lighthearted topic, you know. Uh, happy to help. Yeah. Oh, yeah, of course. Of course. Um, let me just level set here. Does a, I need to be regulated? Don? Oh, I think our answer is, is very similar, but I'll probably present a little bit different than that. You did, which is, um, a, I is a software system. Software system is regulated by rules that have been developed for the last 30 years. They're very robust. They've been part of an ongoing discourse. Um, and so in as much as software is regulated, A I already is regulated and then there's a question of, does it need new regulation? And the answer to that is we'd have to understand what is new about it, which is still a research question. And so until we understand that I would say if you regulated it, you may be regulating the wrong thing if you did it too early. And that could be, you know, even more negative than if you didn't regulate it. So, my view is we have existing robust regulatory regime, it is sufficient for now. And if we understand, you know, that A I is different in some way, we should, we should evolve that, but not until we understand that Dawn. Does that sound like responsible A I? Ok. So I can, I, a lot of that was cribbed by the way, from a lot of that was cribbed from a paper that Don wrote by the way. So I think so. Ok, so, so first let me summarize what is the current lay of land, please do. So if you ask, if you ask this question to the policy makers, they essentially already, you know, speak with their act. So I called this year, we have a certain proliferation of A I belt. So if you look at the federal level, there are about 100 and 20 a belts in flight. And if you look at the state level, there are at least 45 states that are, have A I built in flight about around 600 of them. So it's crazy. So I would say that if you ask this question, there's a madness about A I regulation. Um and going, going back to what my T said, I think there is really um a lot of nuance that we need to look at when we do A I regulation if we don't do it. Well, then for example, if we do it in a way which is mostly what current A I regulation is we have a lot of problems we can have. Um uh we can have a lot of potential unforeseen negative consequences and also even worse, we can actually have a lot of to prevent disasters outcomes. And also we can have fragmentation in community, which is what we have seen unfortunately through, you know, the recent process of some of the bills that will. Um so that's why actually recently we launched a new proposal with a number of leading A I researchers called on the Path for Science and evidence space A I policy, which is excited to address this question. Should we regulate A I or not? How do we even answer this question? So here, then the answer is very simple. It should be science and evidence based. We don't just jump out of the hat whether we should today, we think we should regulate A I tomorrow. We think no, we shouldn't, we should have science and evidence to speak for us. And so that's what the key of the proposal is about is that a policy should be informed by science and evidence and we should do science and evidence based approach for A I policy. And also I strongly believe that this is the best approach forward for us to develop the best A I policy and also the best approach forward to unite the community to reduce recommendation and lead us to a safer A I world. Now, we we are sitting right now in the State of California, we three are residents of the State of California. So let me start at the state level and then we can talk about the federal level if you don't mind. Um, we obviously had SB 1047 come and go as a, a candidly, a flashpoint for the industry was that science and evidence based. Uh, and by the way, that's hardly the only bill that went through uh uh the California. So that's a very good question. So that's actually the reason, I mean, now the reason, but that's one of the motivation is from what we have seen in the process again, like I said, uh so currently most of this A I regulation effort is quite a hoc you can talk about right, the different thresholds. But a lot of these things, we need better, more rigorous approach that gives us the science and evidence and also to really analyze. So, for example, in the proposal that I mentioned, we actually proposed five priorities for A I policy and we propose how you should actually analyze A I risk, it should be risk. I actually I wrote them down. Don it's understand risks, increase transparency, develop early warning mechanisms, develop technical mitigation and defense mechanisms and trust to reduce fragmentation. Those are the five of the proposal, right? Great. Thank you. So, so I think I see SB 1047 is a great example for us to talk about because you know, there are so many policies. So why is this the one that got people up in arms? Right? And so I'll tell you why, which is um again, we have 30 years of discourse on regulation and policy in tech, right? And so we kind of understand basic principles, we kind of understand the basic paradigms by which we apply them. And SB 1047 was a paradigmatic shift in the following way, which it actually included something called developer liability and what developer liability means, which is like if I'm building a piece of software and somebody else uses it and that causes some catastrophic event that I am liable now in like 40 years or 30 years of tech policy, we've never done this and, and maybe, maybe it's something worth doing with A I maybe, but it's a paradigmatic shift that is divergent from 30 to 40 years of policy. And so we're like, well, wait, you know, time out, like, listen, if you want to do policy fine, we kind of understand policy. This is a discourse that's been going on for a long time. You've got a lot of experts, but you're not listening to any of that, you're doing something entirely new and, and based on, you know, as don, who's a world expert on this would say no evidence. And so I would say, listen, it's a very important problem A I policy. It's a very important problem and like, listen, we can even do spurious policies if like we want to because that's what policymakers do and they're going to do spurious policies, like we're going to definitely see that happen. But, but please, let's not create a new paradigm in policy that we don't understand the adverse effects of without actually understanding the risks. And, and the reason is, is we're going to set a precedent and we may end up like policing entirely the wrong thing. And so again, there's a lot of like, there's a lot of, um, new laws and policy being proposed all of the time, but certain ones of them are, are, are particularly pernicious and this was one of them for exactly this reason. And so it, you know, Governor Newsom vetoed it. So it's, it's, it's a moot point. But where do we go from here, Martin? I mean, do we think that we're actually going to take the big step back and do what you're saying? So, I, I think they are very legitimate efforts that are tangible and near term and like winding their way through the courts, right? Things like copyright, I think is incredibly legitimate and we should understand it. Things like child protections are incredibly gate and we should understand it. And so everybody agrees on this and there's consensus and so we should focus on those. So let's focus on the tangible ones that we know are real. A lot of the SB 147 like the proponents are X risk people, like they're like, oh A I is going to take over the world. So we need to put in these precautions that don't make any sense, right? And so I would say, like, listen, let's like, put aside the science fiction, let's put aside like this kind of existentialist stuff, let's focus on stuff that we know is kind of near term and real and impacts people's day to day lives and, and, and by and large, I think that's actually what is happening. I think that we've kind of regained some sense. It's been five years in this journey. Like we as far as we can tell like this A I stuff is actually pretty safe and like we're actually focusing on tangible practical near term regulation. Um OK, so again, coming back, actually, we uh again in the proposal, we made five priorities and uh and the number one actually is to understand A I risk. And actually in there, we actually propose that we do need to understand A I risk in a very comprehensive way. So there are all the risks that my men, but there are also longer term risks. So I think people need to, right? I think it's helpful for the society to do comprehensive study, to analyze, to understand their risk. By now, I think as a as I mentioned that we want to be science and evidence based. So the first step is that we need to better understand the science and we need to collect the evidence. And so I think it is important that we study a wide ranging different types of risks, but we need to understand them. And then the policy needs to be essentially need to be appropriate to the, to the level of understanding that, that we have. I, I think it's actually, um so a lot of people that have been following this don't know how remarkable it is that dawns up here. And so one thing about like these policy wars in the past like 20 years ago is like the forefront was academia and hobbyists and we had these robust debates. And what's been very remarkable about the A I one is like when SB 104 and seven came, like, like academia was largely silent or hadn't weighed in and so forth. And so I think another thing that's been fantastic about this first wave of policy. It is start of the conversation and was brought in stakeholders that really, I would say are far more neutral than say myself, I'm a venture capitalist, right? So clearly, I've got a bias, right? So I should be a voice, but I should not drive the conversation though. In that case, I actually was driving the conversation, right? So I think we're now at a great place where we've got people like Don who are a real expert who are like, hey, listen, like this is a very important problem like, and I think academia is engaged and I think that's actually where we should start moving these discussions to. And so the very concrete next step, let's all fund don this out for us pro pro I get it. So, so let's um, I, I do want to come to the audience for questions in a minute, but II, I want to make the jump up to the federal level. OK. We're gonna have a change in the White House very imminently. The, the outgoing president outlined an A I framework. The incoming one is certainly planning to do just that. How are we going to stick to our science and evidence based guns in this environment? Don. Um That's a very good question, I think especially in the changing environment. I think it's even more important that we focus on like the science, evidence based approach. And uh in fact, we are discussing at all levels um including the federal level and even including like the National Academy and, and uh so, so far, actually, this proposal has received a huge positive feedback. We uh it's only last maybe about two months that we already have volunteer contributors from about 200 different institutions actually um volunteer to help uh with this uh with this effort. And also, yeah, we have received like a lot of positive huge support from civil society from a lot of different, different entities and so on. So we really hope that we can bring this as a whole community society effort. I think it's hugely important. It sounds like a plan. All right, let's go to the audience. Nobody knows what's going to happen under Trump when it comes to like a policy. It's like, nobody knows. Right? Like, I don't know, you know, and so like, on one hand, I'm cautiously optimistic because like what happened under Biden was terrible. It was like he had this executive order that made no sense. And then, you know, it was like this kind of all this kind of crazy A I policy stuff on the other side. Like let's be honest, like SB 1047, which I think is probably the worst, like probably the worst bill, I would say the United States personally because of the paradigmatic shift was supported by Elon, right? And so anybody that thinks that they know it's going to happen, I don't think that they really understand the players, it could go either way. So I'm cautiously optimistic. I think it's good to have a change. I think where we're coming from was actually very bad when it came to A I policy, but no idea, you know what it will actually look like. All right. That's, yeah. Yeah. Yeah, we should ask an A I agent about it. All right, let's let's go to the audience for questions, comments about regulation. And policy. Oh, we've stunned them. Look at that. Yes. Right over here, sir. Your name and who you're with? Uh John Durkan, chief technologist with ID A Ireland. Uh So I'd be very interested from the panel's uh point of view, like particularly looking at the EU A I act that came out recently and the perspective that that has and we saw with GDPR how that kind of set regulation around the world, but in the context of the EU A I Act, your kind of feelings on that, has it been too much focused on regulation or is it because it's took a risk based approach? And it's kind of looking at the idea of setting kind of this idea of being more based on research to make the informed decisions as it goes through. So just interested in your view, I know my answer, but I'll let you go. Um I think, right, I mean, eu in general, they do take a lead a lot of these regulatory efforts. Um and now actually still they have this code of practice after it's ongoing. So I think a lot of details are actually still in the, in the progress to figure out exactly like what the company needs to do to comply, what are the recommendations and so on. And I think they are taking a first step and I think it's leading in the, I mean, certainly it's helping making progress in the space but of course, as I mentioned earlier, there's a lot that we actually really do need to study, to figure out very diplomatic. So I, I come from a very, very poor part of Spain actually. Like, I've got it on the DM, which is like the, the Appalachia of Spain. I love it. It's a great place. Um, you know, uh, we have Alcaraz as a tennis player. He's great. We also have a phenomenal founder who started actually a great company in Ja I um and the company is called Manic. His name is Javier and he's done all of this great work. And today as we speak, he's back in Murcia, you know, this, this, you know, it's great part of Spain but like not known for its tech founders and entrepreneurs. Um and he, and, and by the way, magnific is in the gen A I pixel space. So it's like images and videos and stuff like that and he cannot access Sora. So Sora is the video model that was just released by Open A I yesterday, right? And so like you tell me, but from my standpoint, like when you have like a a, you know, a great founder who can't even engage in his own area because you can't access to something that like a company that Open air release. I think that's a strictly bad. So I think that the eu is totally bundling it. I think it's putting itself backwards and I think it's got to change its tune or it's going to get in trouble. But how do you really feel that more questions, please? Yes, we have one way in the back. Your name and who you're with? Hi, I'm Alana Con, uh, with, uh, hacker one and Admission. I'm a former federal regulator. So I understand the instinct to say we don't need anything. We're already regulated. We've got this unlock. Um, but the federal regulation process, specifically the rule making process, it does take science and evidence into consideration because of course, there has to be notice and comment and you have to be able to provide that information which they take into consideration when they do those rules in the absence while we're off here saying we don't need anything. The Eu Singapore, everyone else is essentially regulating for us. So like privacy, we don't have anything here in the US. We've been fighting it for years and so the eu essentially runs our regulation. So is that the right approach from your perspective? And it doesn't make sense to continue to say we don't need anything and we'll let all these other countries do it for us. Great question. Thank you. Um So first, I don't think we are saying that we don't need anything right? I mean, the the proposal for a science and evidence based approach for a policy does not mean that we don't need anything. Actually, in fact, in the proposal, we actually talk about certain priorities for regulation, including for example, transparency, average event reporting. So we can write to like for post deployment harm and do this average event reporting and so on. So nowhere we say that no means no, no regulation. I think it's more just providing a more grounded uh approach that as grounded in science and evidence. And also again, like I said, and also actually in the end, we actually have um additional cultural action proposal is actually to have the community come together to discuss and propose what we call the candidates um air policies for we call the conditional response so that the community can discuss and what uh conditions, what policy would be appropriate. It's just in this way that we can have um you have multi hold open dialogue discussion in a low stake environment. So we hope we hope that this approach actually can lead to better consensus building process for better policy. Understood, Martin, can I take a quick, I really appreciate this question because I do think it gets to the heart of it. So um I think that our posture as a nation shouldn't change and it's changing. And I want to give you an analog which is like the internet. So I was in the national lab during basically the advent of the web and the internet, right? And we had this new stuff and like very quickly, like it was clear this was marginally different. We had computer worms, which we didn't have before. We had this notion of asymmetry, which means we are more vulnerable than countries that didn't use it. So clearly, it was marginally different, which we don't have for a I. So what do we do as a nation? We invested heavily in understanding it and building out capability. Like we became the leaders of the internet. We turn it into a discipline and like, and yes, of course, there's an entire cybersecurity. You know, like I've known Don for 20 years, he's like one of the preeminence researchers in cybersecurity. So it's not saying that we ignore it and ignore policy since we've become the leaders, we understand it the best and we make the most practical solutions that's not what's happening with A I I mean, the eu is doing whatever it's doing right. It's like, oh, there's this new thing we're regulate it and for us, for me, which has been so shocking is like academia wasn't even involved in the discussions. They've been silent. And so I'm saying yes, we need to regulate, yes, we need to understand, but we need to change our posture to engage research, invest understand the marginal differences. And then yes, let's do like we've done in the past, which I think is the right thing for the country to wrap this up and thank you for that to wrap this up. How are we feeling about where we go from here in in, in uh in one or two words, yours is cautiously optimistic. Yeah, Dawn Yes. So we need to focus on science and evidence and we need to advance our scientific understanding. Got it. All right. A round of applause, please. Thank you very much. I appreciate it. Thank you. Thank you. Thanks folks.