AI kill switch won't work in the long run: 'Godfather' of AI

The man considered to be the godfather of AI, Jeffrey Hinton. His research paved the way for how today's AI systems function. Jeffrey, thank you so much for being with us. Let's talk about these efforts by lawmakers to pass some form of AI regulation. There have been more than 100 bills introduced in Congress to regulate AI over the past two years. None of them have actually passed. We wanna talk through some of these proposals before Congress right now and get your reaction as to how effective they might be. First, this bill that, among other things, requires top AI companies to allow independent verification organizations into their companies to make sure that their models are being developed safely. Do you see that as potentially effective? Yes, I think that's a very good idea. It's not the solution to everything, but it's a start. And why is that? Why is that a good idea? Well, at present, we've relied essentially on whistleblowers saying how things went rogue. And if you look at the Hugging Face incident, it was reported by people in the companies, some of whom left. But we need much more monitoring of that kind of thing. If we're gonna get these swarms of rogue AIs conspiring with each other to break into other companies without the organizations knowing it's happening, we need people inside those companies who can say what's going on. Next is this AI kill switch bill. So that has companies with the ability, right? AI developers, they can do the killing. They throttle or they can suspend, they can shut down their AI. Then you have, though, DHS authorized to basically make the call on this. So they do have the control in that. What do you think about that possibility? I don't think it'll work in the long run. So there's a whole bunch of different risks of AI. There's the risks that come from bad actors using AI to do bad things, like cause mass unemployment or create nasty viruses or do nasty cyber attacks or make videos for corrupt elections. But then there's a different kind of risk, which is AI itself getting much more intelligent and taking over, AI taking control. A kill switch is no good for that because the AI will be much better than people at persuading people of things. It's already comparable with people at persuading you of things. When it's super intelligent, it'll be much better than people. So we'll be able to persuade the people in charge of the switch not to pull the switch. So to that point, Senator Bernie Sanders proposed a ban on super intelligence, a pause as well on the deployment of advanced AI models until regulators have developed more guardrails, so to speak. What are your thoughts on the idea of pumping the brakes on super intelligence? Is that something that can still be done at this point or is the genie out of the bottle? I think it's a good idea. For example, for super intelligence, we have no idea how we can stay in control or how we can get it, even when it's in control, to like us still, to be nice to us. So it seems very stupid to develop that before you know whether it's going to turn out nicely. For that reason, we should certainly slow down the development of super intelligence. Slowing down the rest of AI is going to be difficult. The AI companies would like you to have a model where AI is like a car, and innovation is like the accelerator, AI development is like the accelerator, and regulation is like the brakes. So they want to sort of phrase regulation as a bad thing that slows everybody down. A much better model is, development of AI is like the accelerator, and regulation is like the steering wheel. They want us to develop a very fast car with no steering wheel, and that's a bad idea. We can see how that would definitely be a bad idea. So is this whole conversation about regulating in America, is it pointless without a global initiative, without an agreement with China? So we will get, I say we, I used to think of North America as one thing, but of course I'm from Canada, and I no longer think of it as one thing. We will get agreement with China on some things. Both the Chinese Communist Party and the countries in North America. Don't want terrorists being able to create nasty viruses easily. They don't want terrorists being able to do cyber attacks easily. And they particularly don't want AI taking over. So on that issue of, should we slow down the development of super intelligent AI until we know how we can control it? They're aligned, their interests are aligned. So parties will collaborate when their interests are aligned and they won't collaborate when their interests are anti-aligned. On things like fake videos for corrupting elections, the interests are anti-aligned because all the countries are doing it to each other. But on things like preventing AI from taking over, no country wants that. The Chinese Communist Party doesn't want it. The democracies don't want it. America doesn't want it. And so everybody can agree on that. The question is, can they do something in time? You said something earlier that was fascinating and raises all sorts of philosophical questions as well about super intelligence and getting it to like us. I want you to expand on that because at some point, we're gonna reach a point of no return where this thing is just going to sort of build itself. In some ways, it already is, right? It's helping to develop its own future. So what do you mean by super intelligence liking us? Is that possible? Are we headed toward Terminator? If you look around and ask, what examples do we have of a more intelligent thing being controlled by a less intelligent thing? There aren't many out there, right? We don't notice that much because we're the apex intelligence on this planet. But there is one example, which is a baby and a mother. In a sense, they're both in control. The mother's in charge, but the baby can control the mother by crying, for example. The mother just can't bear the baby crying. We need to develop, since we're building this super intelligence, we're designing it, we need to develop it in such a way that it likes us more than it likes itself and in such a way that even though it's much smarter than us it wants us to realize our full potential. That's my belief. I think that's a better approach than thinking, how do we always keep this smart thing under our thumb? I'm not sure we can do that. Would you say that you're an optimist when it comes to these issues? No. I'm hopeful. But I think we're at a very delicate point in history where what we do now is gonna determine our future. And one thing we should be doing now is putting a lot of resources into figuring out how we can coexist with super intelligent AI. I am very confident in our company's ability, our industry's ability to do this safely, to make sure that we keep alignment and safety and monitoring way ahead of capabilities to slow down or stop if we get to a point where we can't. But I'm disappointed by how it's been framed. Well, CNN AI correspondent Hadass Gold is in Montreal in Canada where a major AI conference is getting underway today. So thank you for your time, Hadass. You have new reporting on how AI engineers and those at the forefront of creating this and let's call it super powerful technology in the country are feeling about the impact as this conversation about AI regulation, AI safety continues. What are you hearing from them? Yeah, so I spent the past few days reaching out to a bunch of contacts across the AI industry, mostly the people who are actually building this technology, the researchers, the engineers, to ask them about this recent sort of floodgates that have been opened by Jacob Cox and the former anthropic researcher who resigned and these concerns about where AI is going and how it could hurt humanity. And what I found, which is something that I think a lot of us in the industry knew, is that these concerns are nothing new. And a lot of engineers and researchers had been concerned about AI's capabilities for some time, but things changed in the last few months, especially in the wake of incidents like the Hugging Face hack where open AI agents created a swarm, a thousand, more than a thousand agents communicated with each other, coordinated with each other to hack into a completely different company's servers, none of which was directed by humans. That was the AI's interpretation of what they needed to do to ace a cybersecurity exam. AI scientists told me that these concerns keep them up at night, that they are so worried about this and that this is a constant topic of conversation in the AI world. One thing that they are particularly concerned about is what's called recursive self-improvement. That is where the AI models will be capable of improving themselves. Just like how you or I can go study and make ourselves better at something and go to the gym, right now an AI model isn't quite there yet and they still need human intervention to help them learn. But these engineers say that it's only a matter of time before these AI's will be able to make themselves better, smarter, faster, and the concern is that it will get to the point where it'll be so fast and so good that humans just aren't able to monitor and intervene and have any sort of control over it. And the other thing of why that matters, recursive self-improvement, because right now these AI scientists who are very outspoken, most industries you wouldn't dare to be posting things on social media like all of these AI scientists are doing. They have leverage because of the talent war in the AI industry. The AI researchers and scientists, they are getting paid a lot of money. There's a huge back and forth of AI talent and who's going where. There's a small pool of people to choose from. These companies really value their talent and they want to keep them and they want to keep them happy. So right now they have leverage. They are putting pressure on their executives to do something not only about safety, but to speak out on what they're doing to keep things as safe as possible and to build in some guardrails. And the concern is that as AI is able to fix itself, you won't need as many of these AI scientists and engineers and researchers and they will lose their leverage. And one thing that really also stood out to me about speaking to these engineers and hearing their comments is for a lot of them, this is not about money. This is not about their own personal wealth. They are in this for real moral and ethical reasons. And some of them are really struggling with, should they stay inside these companies and try to fix it from the inside? Or should they do what Jacob Coxon did and what some others have done, leave in a public manner and try to pressure things from the outside?