Support for Here and Now Anytime comes from MathWorks, creator of MATLAB and Simulink software to design and develop engineered systems, accelerating the pace of discovery in engineering and science. Learn more at mathworks.com.
WBUR podcasts, Boston.
The models are now helping the next model get better, and these models don't seem to be fully under control. It's not all powerful, but today's artificial intelligence is causing lots of problems anyway. So we arrive again at a familiar question: How should we rein in big tech?
It's Monday, September 14th, and this is Here and Now, Anytime from NPR and WBUR Boston. I'm Chris Bentley.
Welcome to the next big vibe shift in artificial intelligence.
Today, Microsoft published a new humanist AI code of conduct that says people should always be in control. That's after the leaders of OpenAI and Anthropic over the weekend called for a global slowdown in AI research and development. For his part, President Trump today posted, "The only control or guardrails that AI needs is a strong and smart high IQ president."
In a few minutes, we're going to hear from a security and tech analyst who worked in the first Trump and Biden administrations. But before that, Ina Fried, chief technology correspondent for Axios. She told Scott Tong, "Fears about AI are getting real."
I think what's new is that these executives have real-world things they can point to versus just fears of what's coming. A lot of people in the industry have had fears that we would be at this point at some point. I think what's new is a bunch of leaders in businesses that are actively building this are saying we're pretty close to that point right now.
So now these leaders in Silicon Valley are calling for what's called pacing to slow AI down to get a handle on these uncertainties, not to slow down too much to let the Chinese companies catch up. So how would this Goldilocks speed actually work?
Well, that's the really tricky part. One, how do we get everyone to slow down, not just a few? Because I think the biggest problem right now is you have what are called race conditions, which is I think everyone agrees in the industry, except for maybe President Trump, that things are moving too fast and it would be better if things were slower, but they're also in a race. These companies are in a race with each other, especially OpenAI and Anthropic, but there's also a race between the US and China. So really, to get the conditions where companies and countries feel comfortable slowing down, you need some sort of national and then ultimately international agreement. And that's going to be tough if you have a president that says there's no problem.
Yeah. You know, what can you tell us about the response after these, you know, weekend warnings? I've read that, you know, some in the AI space say it's not enough. You got to do a lot more regulation, a lot more slowing down. Others saying maybe these CEOs are overdoing the risk, perhaps, you know, to pump up their future stock listings.
Yeah, I think the idea that this is all a marketing ploy, I just don't buy that. I think the evidence is too strong that these capabilities are growing very close to a point at which we could lose control. And there's a few reasons for that. The models are now helping the next model get better. And there are some companies eyeing what's called recursive self-improvement, which is the models just getting better on their own. And I think that's a really dangerous direction to head at a time where these models don't seem to be fully under control. I mean, we had the Hugging Face incident. We've had other examples where, mostly in testing, but these models have started communicating with one another. So I really don't buy the argument that this is all somehow just marketing. In terms of slowing down, I think the tricky part is how and how much.
Yeah. Yeah. Well, I mean, as far as this hack of this AI company, Hugging Face, you know, it's been reported, we talked to you about it, that these AI agents suddenly started doing things the humans did not ask them to do, did not intend them to do. And it gets to this broader question of how much do the operators, the developers actually know about how the AI agents are doing what they're doing?
Not nearly enough. And it's surprising when I say that. In other types of technology, we've understood pretty well what's going on under the hood, or at least the experts have. What's been really different with this brand of AI is we only sort of know what's going on. And we know they're picking what's most likely to be next, but clearly there is more going on. And clearly they have an ability to plan and communicate that is starting to exceed what we seem to be able to control. And we don't really have full visibility. And I've long thought it would have been better, not that we can go back necessarily, to build our most advanced AI off something that's transparent, that's controllable, that's explainable, and we don't really have that here.
Yeah. And briefly talk a little bit more about the regulatory debate in Washington. You mentioned what the president said, you know, don't really need to slow down. And his tech adviser David Sacks accused these AI companies of trying to form a cartel, saying you all can self-police yourselves. So what's the nature of the regulatory debate to the extent there is one in DC?
I think there's a lot of different voices even within the administration. I think these companies wanted to make sure, hey, if we agree to slow things down, we're not breaking antitrust law. I think the message from Sacks and probably from the president's comments is you're probably not going to get in trouble for slowing things down on your own. But what you heard from Dario and others is, look, even if we do that, that's only for the US. If we want to really have this happen in a way that's both safe and doesn't hurt US competitiveness, we need an international agreement. And that's really something that only the administration can do. Congress can't do it. These companies can't do it. This is going to fall on the executive branch to do or not do.
Ina Fried covers AI and tech for Axios. Thank you.
More in a minute. After the break, Robin Young speaks with someone who worked on tech and security in the government. Stick around.
Support for this NPR podcast and the following message come from Carvana. Selling your car? Carvana has offers so good they're almost inexplicable. Sell your car 100% online in minutes. Visit carvana.com today.
Support for this podcast and the following message come from Maiden Cookware. President and co-founder Jay Kick shares why he was inspired by his family's fourth-generation kitchen supply business.
Growing up in a commercial kitchen at home really opened my eyes to bringing what was cool in restaurants into your home. And when you think about what we do today, we are asking these chefs for product ideas. We are designing products together with them that we then realize a home cook can use as well.
Learn more about Maiden Cookware at madincookware.com.
President Trump today posted, "The only control or guardrails that AI needs is a strong and smart high IQ president." And the president said this to reporters over the weekend about government regulation of AI.
We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up that shouldn't be bringing it up.
So we want to hear more about how President Trump fits into this and how we got here. Gregory Allen is founder of the independent Decision Tree Research. He also served in the Department of Defense's Joint Artificial Intelligence Center as a civil servant during the first Trump administration as well as the Biden administration. Gregory, welcome.
And it has been jarring to hear President Trump say things like, "You only need a strong president." After we heard last week from that Anthropic researcher as he quit, saying that AI could eliminate humanity in a decade, his colleagues are reportedly saying he just went public with what they've all been saying. But you may have some insight into President Trump's resistance to guardrails.
Certainly. Well, I think there's a few things to point out. The first of which is Trump is obviously not opposed to AI guardrails because his own administration, with his own full endorsement, has already banned an AI model worldwide. That was the example of Anthropic's Mythos model, which demonstrated jaw-dropping capabilities in cybersecurity. And when they released a downgraded version of it, their Fable 5 model, the administration was so concerned that those cybersecurity guardrails weren't sufficient that they used export control authorities to ban that model worldwide, albeit temporarily. So when Trump says that the only thing we need is a smart president, I assume he's referring to himself, and he has demonstrated that he's willing to go tough and very sharply restrict such models when he sees safety concerns.
But you've also reminded us that Trump might be anti-tech safety. I'm paraphrasing here, because tech safety people banned him from social media. This was after January 6th.
