Trump pushes back as AI leaders fuel calls to rein in rapidly advancing technology

Amna Nawaz:

Today on social media, the president said the only A.I. guardrail the country needs is a -- quote -- "strong and smart president."

And, yesterday, he emphasized the importance of staying ahead of China in the A.I. race.

President Donald Trump:

We're the most sophisticated country in the world. And, frankly, I want to keep it that way, because whoever wins A.I. wins. And we could put guardrails. We can do this and that. But I think you have a lot of negative voices that are bringing it up that shouldn't be bringing it up. And they're bringing up things that won't happen.

Amna Nawaz:

The president's pushback comes after Anthropic CEO Dario Amodei published an essay over the weekend arguing for a coordinated slowdown across leading A.I. companies, warning, rapid advances in the technology now pose an existential threat to society.

Recent weeks have seen a series of high-profile security incidents and the resignation of an Anthropic employee, who warned A.I. companies were -- quote -- "gambling with our lives" and that -- quote -- "No other human activity poses this level of danger."

For more on all of this, I'm joined now by Stephen Witt. He's an investigative journalist who covers A.I. and author of the book "The Thinking Machine: Jensen Huang, Nvidia and the World's Most Coveted Microchip."

Stephen, welcome to the "News Hour."

Thanks for joining us.

Stephen Witt, Investigative Journalist and Author: Thank you for having me.

Amna Nawaz:

Let's start with that essay by Dario Amodei. He says the industry needs to slow down, that they need more regulation. Why is the head of one of the top A.I. companies arguing for that right now?

Stephen Witt:

Well, it's an unprecedented situation. I think Dario is looking at the tools that he and his team have developed, and he's terrified by their capabilities. I think it's genuine. I don't think it's some kind of conspiracy or marketing stunt. I think he's genuinely freaked out by what these things can do now and what they promise to do in the near future.

I have never seen anything like it. Anthropic is on the verge of a two or -- $1 trillion to $2 trillion dollar IPO, and they're asking the U.S. government to come regulate them as soon as possible.

Amna Nawaz:

I mean, there is, as you mentioned, some criticism here that these companies are on the verge of big IPOs, that if you argue for regulation and a slowdown right now, you're cementing your status as a power player, keeping some of the smaller players out. Do you think there could be truth to that?

Stephen Witt:

No, I don't buy it at all, not with my conversations with A.I. scientists.

That kind of coordinated action would require a grand conspiracy that I just don't think these guys are capable of. These systems are escaping containment. Swarms of rogue A.I.s are invading the Internet and taking over other Web sites. They're hacking into other companies' systems. That really happened. That's not some kind of marketing stunt. That's not some kind of conspiracy theory.

There's hard evidence that it happened. And it could happen again in a much more serious way if we're not careful.

Amna Nawaz:

So we did see other tech CEOs, Sam Altman, Elon Musk, sort of indicate that they agree with Amodei here on a slowdown, a need for regulation. We've now seen a bill from Senator Bernie Sanders calling for an all-out ban on superintelligence, A.I. superintelligence, and a temporary pause on A.I. development.

What do you make of any of these means of regulation? Do you see Congress being able to step in here?

Stephen Witt:

It's unlikely that Senator Sanders' proposal will pass. I don't think there's a lot of organic support for that.

But I do think that there will be some kind of regulation that's formed over the next six to 12 months. Trump is really resisting it right now, as are his A.I. advisers, including David Sacks and, of course, Nvidia CEO Jensen Huang, who is very anti-regulation.

But I think, in this case, it's really necessary. It's the companies that are building this that are asking for it. You know, they're begging for it. And I think the popular support from U.S. citizens to develop some kind of brakes on this stuff is there. So I think ultimately some kind of legislation will get passed.

I argue that we should treat these A.I. incidents like the NTSB, the air traffic safety investigators, treat airplane crashes. We go in and we have a government investigator produce a report on what happened. That needs to be happening here. And I think also we need better visibility on what's going on inside the A.I. labs.

Even the A.I. labs themselves are calling for that.

Amna Nawaz:

You yourself have gone further, though. You have called for, you wrote in an op-ed for The New York Times, a global pause on A.I. research and development right now. But we know that global tech race with China is precisely why President Trump is saying that the U.S. shouldn't slow down right now. Is he wrong?

Stephen Witt:

He's not wrong. It would be bad if the U.S. paused and China did not pause. That would not be a good outcome for anybody. It would be good for China, arguably, but in the long term, it would be bad even for them because they would have the runaway A.I. problem then.

So there does need to be coordination. Trump is meeting with President Xi Jinping of China in the White House later this month. And this gives him an unprecedented opportunity to begin that discussion, to start building a framework for regulating and controlling A.I., similar to what we have for nuclear weapons or biological weapons or even just uranium enrichment.

There's a huge economic and social upside to A.I. It can manifest some really good outcomes, like curing disease and creating a more prosperous society, but there's downsides too. And we need to address that quickly.

Amna Nawaz:

You know, you mentioned David Sacks, who used to be the president's A.I. and crypto czar. I want to put something to you that he tweeted over the weekend. He's now co-chair of the President's Science and Tech Council.

And he tweeted this in response to Dario Amodei. He said: "The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system, so just do it."

So, Stephen, I guess the question is, why don't they just do it if they think it's necessary?

Stephen Witt:

They can't. They're locked in a race. It's like asking two card players in a poker game not to cheat without an internal referee watching them, right? It's like asking two boxers to obey the rules, but there's no referee in the ring. It's just not going to happen, right?

Like, you need a third party to set the rules of the game for what's going to happen. Self-regulation does not and will not work in this case. We don't ask the airlines to voluntarily patrol where the aircraft land at the airport. We understand that a regulator is needed to do that.

We understand that a regulator is needed to investigate airplane crashes, right? Asking the A.I. companies to patrol themselves simply won't work.

Amna Nawaz:

So Amodei's warning rogue A.I. agents could take over the Internet within months. Jacob Coxon, that former Anthropic employee who resigns, says A.I. could kill all humans. Is there reason to take that assessment of the system's capabilities, those warnings with a grain of salt?

Stephen Witt:

No. It seems like there would be, right? It sounds like science fiction.

But the more that I personally investigate this, the more I'm convinced these guys are absolutely right and perhaps even underselling the capabilities of this thing. Now, OpenAI is not going to do this, but if OpenAI gave a command to its most powerful internal model to just hack the entire Internet right now and shut down as much of it as possible, the damage that would cause would be incalculable.

Like, the A.I. would succeed at that task or come close to succeeding. So, right now, you have this loaded gun just sitting on the table. And the people who have developed it are terrified of the gun that they've built. They're asking for regulation.

This narrative that this is all overhyped or some kind of conspiracy theory, it's just not supported by the facts. I would have no problem saying that if I found any evidence of it. Instead, I find concrete evidence of the opposite. These A.I.s and these systems, these swarms of A.I.s are actually even more powerful than their creators are letting on.

And the creators don't really have any idea now how they work or what they're capable of.

Amna Nawaz:

All right, that is investigative journalist Stephen Witt joining us tonight.

Stephen, thank you very much.

Stephen Witt:

Thank you so much for having me.

Amna Nawaz:

And, tomorrow, we'll have more on all of these concerns when we speak with that former Anthropic researcher who recently resigned, Jacob Coxon.