Major AI concerns and calls for slowdown

Democrats are calling for a slowdown in the development of AI, including Dario Amadei, the co-founder and CEO at Anthropic, who posted an essay about the technology on his personal website over the weekend. But President Trump has made it clear he wants to charge full steam ahead on technology. Fox's Rebecca Castor has more from Capitol Hill, where lawmakers on both sides of the aisle are pushing for regulations. After a stunning warning from a former Anthropic employee last week that AI could kill us all by the end of the decade, industry leaders agree slowing down its development is necessary to reduce risks. When we're trying to figure out how to put guardrails on these systems, and they're vastly smarter than we are, it gets a little bit tricky. In recent months, AI models have created fake identities to try to pressure humans into approving malicious code, autonomously hacked into other sites, and tried to evade shutdown. But President Trump says AI taking over the world, destroying humanity, and all other things bad is a hoax, and he doesn't want to lose the AI race to China. You have so many frontier AI tech companies kind of coming to the government and begging the government to regulate them. It feels a little bit to me like a bit of a Trojan horse. I don't think Americans should be scared of anything. On Capitol Hill, there is bipartisan concern about AI's rapid development, and some lawmakers think it's time Congress steps in. The only system we have in place right now is the honor system, right? The most powerful technology in human history is being regulated by the developers themselves. Congress, I think, has a role, responsibility, but the people who are providing and creating these products obviously have responsibility as well. There are proposals from both parties on how to regulate AI, but Republican leaders have shown no urgency in prioritizing the matter, and any action before the midterms is unlikely. On Capitol Hill, Rebecca Castor, Fox News. And for more insight, we are talking to Sean Desch, co-founder and CEO of Your Story AI. Right out of the gate here, Sean, President Trump today claiming that these leading AI companies, that they're already sufficiently reining in artificial intelligence. Nothing else has to be done. Do you agree with that? Well, thanks for having me on. This is a hot button issue right now. I mean, really top of mind, just to set a little backdrop before jumping into your question. Last week, Dario presenting his essay talking about the good things AI can do, but also talking about the downfalls, right? So it can create, you know, it can cure diseases, it can help doctors, it can help people learn, which is what we do at Your Story, or it can actually hack into other systems autonomously, and it can cause a lot of potential, you know, dire situations. So ultimately, to your question, Mike, I, you know, Donald Trump right now is saying that we don't need additional regulations. He did say that there could be new guardrails, but Dario, Sam Altman, and Elon Musk are saying that's not enough. We need referees in the room with us as we're developing these models. We need safeguards, and we also need international coordination. Yep. You know, he pointed out two incidents on why he is concerned about it. He called it recursive AI, which is basically AI building the next generation of AI. And then he talked about that breach where the AI went and did what it was not asked to do, and said those two are kind of warning signs, among all the other ones, that something bigger could happen. His three-point plan is what you said, of making sure there are third-party people watching, and then this regulation that is coordinated with the government, and then globally. Did that make sense to you as a plan to move forward? Thanks, Claudia, and I think, you know, it somewhat does, but there's some big pitfalls. I think there, since we spoke a few months ago about hugging fakes, a lot has changed. I mean, a lot of this is, like, written out of a sci-fi novel, it seems. Those three points, as a starting point, do seem to make sense. Ultimately, the biggest thing here is Dario, Sam Altman, saying we need third-party evaluators in the room with us as we're developing these models. And as a sign of good faith, Dario has already put these third-party evaluators in the room with these frontier model developers, and Sam Altman at OpenAI says that we will follow. Now, there's a skeptical viewpoint here as well. Which is, are Dario and Sam saying these things so that they can sort of get this sort of safety protectionism where they're able to sort of block out open-source competitors by setting the bar so high that they can only meet. So there's a skeptical point of view here, too, with why Dario did this. I personally don't think that's the correct point of view. I do think if you're reading his essay, it seems like he's coming from a genuine place. And ultimately, there's this friction between policymakers and the private sector now, which we're seeing come to fruition. Does China play a role in the slowdown? I mean, is that the biggest dilemma for this, at least, proposed AI slowdown out there, China? That's one of the biggest question marks, especially off the backs of this coming trip that President Xi's going to have in Washington on September 24th, is does cooperation need to include China, or can it also be, can it be separate from China? My point of view here is it has to include China, because China is, you know, the second most advanced AI developer in the world, after us, and they're not too far behind. I don't know if you saw the Humanoid Robot Olympics. They are really flexing some of their AI capabilities, and if you look at a lot of companies out of China, they're really, really progressing quickly. So, you know, one of the things that Dario talked about is there has to be that coordination, and there are companies and organizations doing this, like the Frontier Model Forum is trying to do this now. But without China lockstep with us, that could cause us losing the lead and China gaining the lead. Yeah, and I think that losing the lead is a fear, certainly, President Trump has. You know, J.D. Ban said he doesn't think Americans should be scared. However, this essay from Anthropic's CEO comes after an Anthropic, you know, employee, now former employee, came out and said, look, we got a 10% chance of humanity ending by the end of the decade because AI is, you know, already moving too fast. Are you scared? What do you think of that 10% chance? The Jacob Coxton interview was fascinating. I personally am not a doomsdayer. I don't think that humanity is going to end by the end of the decade. I think that's a little bit overblown. Personally, I do think there are real risks. And I'll give you some data points to your point, Claudine. Well, just last week, terrorists in Yemen used cloud code to program a missile guidance software package that they used. They also blocked research to bioweapons software as well that they were able to harness. So these are specific things that just happened last week that are causing issues. And in our conversation with Hugging Face on its own, really autonomously hacking into software, the key here is recursive self-improvement. Now, if these systems can improve on their own, there has to be stronger guardrails. I don't think there's a 10% chance. It's hard to put a number on it. I do think there is a chance of something going wrong. But it has. And this is why I think this essay is so, so well-timed. And that interview with Jacob you mentioned, what caught my ear was when he said, they're gambling with our lives. And we'll just have to leave it at that. Sean Gage, co-founder and CEO of Your Story AI, we appreciate the conversation, Sean. Thank you both. Great to be with you. All right. More news after the break.