Treasury Sec. Bessent, AI Control, & the White House Press Pool 9/21/26

0:00You spent millions deploying AI across your organization, but six months in, it still misses basic context. Sound familiar? That's not an AI problem, that's a data problem. Siloed systems leave your AI blind, while dark data confuses whatever it does see. EverPure unifies your data so your model can see it, understand it, and put it to work. No waiting around, no digging for answers. Get real results from your AI at everpuredata.com. This episode is brought to you by Schwab Market Update, an original podcast from Charles Schwab. Join host Keith Lansford for this information-packed daily market preview delivered in 10 minutes or less, including projected stock updates, monetary policy decisions, and key results and statistics that may impact your trading.

0:48Download the latest episode and subscribe at schwab.com slash marketupdatepodcast or find Schwab Market Update wherever you get your podcasts. Bring in show music, please. Hi, I'm CNBC producer Katie Kramer. Today on Squawk Pod. Treasury Secretary Scott Besant on artificial intelligence. Is it time for the frontier labs to pause the technology's development? Then the labs also said, take the liability off of our hands. And we will not do that. His take on the chance AI could kill humans. If you were running a fund and someone said, oh, there's a 10 % chance it's going to be a zero, you probably wouldn't invest.

1:31And the continuing sanctions against Iran. I can't tell you whether it's one week, one month or six months, but I will tell you that we are pressuring them like never before. So another week, another debate on AI ethics. Berkeley professor Stuart Russell has literally written a book about it. The real question is, does this technology path that we've embarked on ever support the level of safety that we need? Can humans maintain control? We don't understand how the technology works. We don't design it. We don't program it. We grow it. Plus, news organizations CNN, MSNOW and Politico to sue the White House over a ban.

2:13Our Eamon Javers has the latest from Washington. Does this show that you are too weak to handle tough questions from a tough press corps? It's Monday, September 21st. Squawk Pod begins right now. Stand Becky by in three, two, one. Cue it, please. Good morning, everybody. Welcome to Squawk Box right here on CNBC. We are live from the Nasdaq market site in Times Square. I'm Becky Quick, along with Joe Kernan. Andrew is out today. President Trump and China's President Xi will hold a summit in Washington Thursday. Treasury Secretary Scott Besson met with China's vice premier here in New York yesterday and said that the meeting was successful.

2:55They discussed a U.S.-China AI dialogue proposing that the country's set up a notification mechanism for AI incidents. The notification for whether there are incidents, whether there has been something that rises up to a national security level from from AI, Because we think that just just like with any cross border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world, very important. We're going to talk a lot more about this. I've got a bunch of AI generated questions for Scott Besson today. They must be really good. You can expect a good interview.

3:48No, that's not true. But anyway, that will be expected. Yeah. CNN, MS Now and Politico all set to file suit over being banned from the White House. Let's get to Eamon Javers, who has the details. Hey, Eamon. Yeah. Good morning to you, Joe. The three media organizations that were banned by President Trump from the White House grounds Friday are filing a lawsuit in federal district court in Washington this morning. They're seeking a temporary restraining order that would reinstate their access here to the White House grounds. The suit, filed jointly by CNN, Politico and MSNOW, argues that the president's actions violate the First Amendment of the Constitution because the president is punishing them as a result of speech that he doesn't like.

4:31In a joint statement, the three news organizations said, This morning we notified the government that we are filing a lawsuit today to protect our First Amendment rights and defend the principle that the government does not decide what the press reports or publishes. Without notice or process, the White House revoked our journalists' credentials because it objected to our reporting. Left unchallenged, this threatens press freedom and the public's right to independent journalism free from government interference. Now, on Friday, shortly after announcing the ban on social media, the president explained why he took action against the three outlets.

5:04Here's what he said. They know that they know it. They know it as well. Well, they're smart people, but they purposely write negative news. And they do that because they want to try and diminish the Republicans and a Republican administration. Over the weekend, reporters from all three outlets were denied admission to the White House and their credentials were confiscated by Secret Service officers here. One challenge ahead for today, guys, is that CNN was scheduled to be the TV pool network covering the president's movements through the day to day. But the White House released a schedule last night that no longer includes that network.

5:41So the behind the scenes of this means that we may not be able to see the president on camera today as he leaves for New York and meets with officials there. And we should, of course, point out that MSNOW has the same parent company as CNBC, which is called Versant Media. Back over to you. Right. I would really hesitate to say there's good people on both sides. But but but just over the weekend, I see so many different I see arguments on both sides on Twitter, as you normally would on on on the one side. It's that this is the people's house. It's not the president's house, the White House. It's it's he's an occupant.

6:17He was elected there. But the press needs open access in. It's a fundamental right of the press to be there. On the other hand, then other people say, well, wait a minute. It's a privilege. It's not a press. It's not granted to everyone. Others said that no other president ever did this. Other presidents did it a different way. He has been probably more open to taking questions nonstop, because I think he likes to do that, versus you remember in President Obama's administration, a lot of these things were private. They would just be called private and there would be no press there. It'd be like I think there'd be about one in 10 opportunities available.

7:02And the press complained about it at that time. Accessibility to President Obama. I think you were doing this back then, weren't you, Amy? Look, I mean, there's always a question about press accessibility and the press always wants more. Right. This is a different issue, though, Joe, because what you have here is the president of the United States banning certain outlets from even coming on to the White House grounds. because he doesn't like the questions he's getting and the stories he's getting from them. That is an attempt by the White House to suppress coverage that they don't like. And it's un-American.

7:33You know, this is an authoritarian style maneuver. And I think, you know, the question politically for the president is, does this show that you are too weak to handle tough questions from a tough press corps? You know, that is part of our tradition as Americans is that our presidents, whether they like it or not, have to handle hostile questions in the press corps. What you've seen this president do over this term, Joe, is slowly weed out outlets that he doesn't like by putting that pressure on and then bringing in outlets that he does like. He's brought in a lot of MAGA influencers and things like that.

8:03Do you remember the record here and to be in the Oval Office with him and ask him favorable questions? Do you remember the record between press conferences for Joe Biden? How long was it? I don't know the number of dates, but there were certainly complaints. Was it a year? I didn't talk to the press enough. Yeah, absolutely. And there was a clear effort by the Biden White House to hide his condition and his inability to his inability to take questions. And that's why the press wants to be here asking questions and having access is because you can then determine if there's some kind of problem with the president.

8:34Right. And the president's inability to take those questions, I think, was revealing in the Biden administration and maybe revealing here as well. Yeah. I mean, we never he didn't have to ban anyone. No one got to ask him. We saw Mitch McConnell in the last three months more than we saw Joe Biden. All right, Eamon, appreciate it. Thanks. You bet. Shares of Paramount moving higher. The Wall Street Journal reports that Paramount and California's attorney general have discussed concessions to settle a lawsuit blocking the studio's$81 billion acquisition of Warner Brothers Discovery. Those concessions include Paramount investing$1.5 billion in the state, a promise not to sell either studio lot and to stay in California.

9:19The party's also considered potential penalties if Paramount doesn't make good on its pledge to make 30 movies a year after the merger. 30 movies a year that actually make their way into theaters, I think was the deal they would have been talking about. A final deal hasn't been reached. Warner Brothers shares also moving higher on this. They're up by about 7.25%. Yeah, that's what happens, I think, when you hear about all this property in Nashville that the company is thinking about. You believe them. That's the question. That they would actually move. You believe that threat. And set it up. Cheese will be next.

9:58Next on Squawk Pod, the AI dilemma with Berkeley, a professor and president of the International Association for Safe and Ethical AI, Stuart Russell. Just because we can do something doesn't mean it's in our best interest. I think it's almost certain that if we continue to invest at the rate we're investing, we will get to machines that are substantially more. I know some people that machines are already small, actually, when I think about it. So what will it take to create a governing body for AI? Plus, comments from NVIDIA CEO Jensen Huang. We'll be right back. If you like YouTube, you'll love YouTube Premium.

10:35It's destroying athlete, creator, and YouTube maxer. YouTube Premium enhances how I use YouTube with awesome features like offline downloads so I can download my favorite training videos before I hit the gym. So no Wi-Fi doesn't turn leg day into loading day. Plus, I get ad-free videos, background play, and so much more. YouTube Premium is like YouTube got some extra gains. Try YouTube Premium for two months free at youtube.com slash premium. Trial eligibility varies. Terms apply. Cancel anytime. time. This episode is brought to you by Schwab Market Update, an original podcast from Charles Schwab.

11:09Join host Keith Lansford for this information-packed daily market preview delivered in 10 minutes or less, including projected stock updates, monetary policy decisions, and key results and statistics that may impact your trading. Download the latest episode and subscribe at schwab.com slash market update podcast or find Schwab Market Update wherever you get your podcasts. When you partner with CDW, you get more from technology with AI-powered tools that turn complexity into clarity. CDW experts are simplifying workflows with Microsoft solutions, uniting teams with on-the-go connectivity, speeding workflows with exceptional performance, and protecting assets with chip-to-cloud security.

11:53Make amazing happen. Learn more at cdw.com slash modernwork.

12:06Welcome back to Squawk Pod from CNBC. NVIDIA CEO Jensen Wong sounding off of what he calls AI doomsday narratives. Wong telling CBS News that he disagrees with the recent warnings that the technology could lead to the demise of the human race. 2030 is not going to be the end of the world. There is zero percent chance that's going to be the end of the world. And so I think that scaring people is unnecessary. It is irresponsible. And I don't know what their motives are. Wong was responding to claims by former Anthropic researcher Jacob Coxon earlier this month that AI developers, in his words, earnestly believe that it could kill us all by the end of the decade.

12:53Anthropic CEO Dario Amode and others have also called for a development slowdown. Wong was also asked why Americans should trust him on AI safety. Our company's success is directly connected to the safe deployment of products and services. If the computer industry doesn't provide value to the world safely and productively, provide enormously positive impact as we have so far, If we don't continue to do that, our value would be diminished. NVIDIA's market cap is roughly$5.3 trillion. That stock right now up by about 1 % this morning to$224.42. I love that, man. But, you know, the questions, the motivations of other people, I guess people would question, you know.

13:49NVIDIA is dependent on the AI build-out, obviously. So you could ask the same questions about Jensen. It's very odd. A couple of stories on 60 Minutes, on the new Barry Weiss 60 Minutes. The data center's one? Just that it's like 100 % negative when there's these town halls. When it comes to your town. And they didn't focus on that Texas town. Right. It would have been 25%. It was an area the size of Central Park in New York City that they were considering for these things. And I think some of the town leaders that they spoke with said, we want them to tell us what they're doing, not just assure us that they're good and they're going to do this and it's going to be excellent for us.

14:29We want to know what's going on. It's the questions that people are asking that they hadn't been asking. I mean, there's so much. I guess you need employees so you can't put it in a complete desolate area. But I just look, there's just there's places where your next door neighbor is 12 miles away. still all across the parts of this country. I think the other issue is that it will bring in tax revenue, and then they said that Amazon Web Services had been asking for 90 % of that to be given back or get the rebates on that for the next 30 years to not be paying those taxes. Look, when Jensen says, I guarantee you it's not going to be the end of the world by 2030, he's probably right.

15:13If he's wrong, nobody's going to be here to tell him he's wrong. But I think what we're doing with this debate at this point is putting it in these crazy extremes, either the end of the earth or everything's perfectly fine. And most people think it's somewhere more in the middle, that this is maybe something that should be regulated a little more, like you have with nuclear energy, like you have even to get a drug out of the FDA. We're going to speak with someone in just a moment. This professor, Professor Stuart Russell, who can tell us a little bit about what the regulatory structure might look like.

15:47Yeah.

15:50I would think it's zero. I would go along with that, that the actual end of the human race. I'd go out on a limb. I think I would take that side of it, too, if you're forced to take two crazy extremes. Now, I don't know about the Al Gore tidal wave hitting New York. That could be well above. Well, no, probably not. Our next guest has been sounding the alarm on AI safety and calling for AI regulation. Stuart Russell is a distinguished professor of electrical engineering and computer sciences at UC Berkeley and the president of the International Association for Safe and Ethical AI. And, Professor, there's a lot I want to talk to you about.

16:30I mean, you have been warning about this for a very long time. you've written the book that all of these colleges now study about learning about AI. You started saying a few years ago that you would not do interviews with people who started using the Terminator imagery in any of these interviews. You don't think that is the likely scenario. So let me ask you, what are your biggest concerns about where we might be headed? So let's put it very simply, right? The companies developing this technology is spending trillions of dollars to create machines that are much more intelligent than human beings.

17:08Those machines will be much more powerful than we are, just as we are much more powerful than all the other species on Earth because of our intelligence. So how do we expect to retain power forever over entities more powerful than ourselves? That's the question. Alan Turing, the founder of computer science, asked that question in 1951, and his conclusion was we would have to expect the machines to take control. How do we prevent that from happening? I think very simply, we don't build machines that have that capability unless we can prove beyond a reasonable doubt that they are actually safe, that they will not take control.

17:56I mean, think about it. What probability of human extinction would you be willing to accept? Well, we accept a risk of about 1 in 10 million per year for a nuclear plant to melt down. So if you want to operate a plant, especially near a city, you have to prove to the regulator that that's your failure rate, one in 10 million per year chance of failure. So obviously, extinction is much worse than that. We probably can't reduce it to zero, but we could say it should be comparable to the background extinction rate from supernovas and gigantic asteroids and things like that. So maybe one in 100 million per year, we would be willing to accept across all AI systems.

18:48What does that look like? I mean, to try and regulate that, we know what it looks like for nuclear energy. You've also compared it to the FDA, which has all kinds of approvals you need to get through before you can introduce a drug onto the market. And I think you've said the same thing about Boeing. If Boeing said that it wanted to introduce an airplane and how did you say it exactly? So I was responding to Dario Amadei's letter called, We must pace the frontier. And I think a lot of people interpreted that as a call for just slowing down, which would give us a bit more time to work on safety.

19:26And I was saying, you know, if Boeing said we're going to slow down, we're only going to introduce a new airplane once every year. And we hope that that gives us enough time to do some flight tests and maybe get some good results. We wouldn't accept that. The FAA would say, no, it doesn't work that way. You have to do the flight tests until you get good results. And then you can introduce the airplane once we certify it. And in fact, this is what Dario is saying and this is what I'm saying. We set the requirements for the acceptable risk level. And we require evidence from the companies that their systems meet that acceptable risk level.

20:09We even do this for sandwiches and hairdressers, right? All the restaurants around Times Square, they have to get permits. They have to convince the authorities that they know how to run the restaurant, that they have hygiene, that they train their employees, that they're sourcing from proper suppliers and so on. So if a little restaurant across the road can do that, I don't see why a trillion dollar company can't provide the necessary evidence to the government and to us. Right. Because we're the ones who bear the risk that their system is actually safe. But what are they actually telling the government?

20:46They're saying there's a good chance, 10 percent, 20 percent, that we are going to kill everyone on Earth. What do you want to do with RSI? Is that, it's not happening yet, right? Not yet. No, it's, I would say we're seeing early signs of losing control. Right, but it's not there. Is it theoretical or is it only a matter of time? Absolutely. Is it 100 % that we'll get there? So I think there's two phases, right? There's will we create machines that are substantially more capable than human beings? Being able, this loop that we think where they're already smart and they're making themselves smarter and smarter and smarter every time they do it.

21:29Is that on the horizon? And don't we have to do something before? Isn't that past the point of no return when they get there? When that loop becomes completely autonomous, then yes, all bets are off. We don't get to say what happens at the end of that loop. Right now, I don't think it's quite fully autonomous. I wonder how many years. And then also AGI. I wonder whether that's real. Is that going to happen? I think it's almost certain that if we continue to invest at the rate we're investing, we will get to machines that are substantially more. I know some people that machines are already small.

22:07Actually, when I think about it, there's plenty of people I've met where machines are already. You get mad at Claude all the time for not giving you the right answers. Claude, what I love about Claude is he may know nothing, but he's never in doubt. He gives me answers that he, until I say, are you sure? And he goes, oh, God, I checked. And what? He apologizes. We call him he, right? But it apologizes. It does. You know, I checked a little more. Right. And so often it apologizes and says, oh, I looked on the web and I couldn't confirm this. And that's a lie as well. Because he didn't. Because he didn't actually look on the web.

22:46But you think AGI is a real thing someday. I think so. And I would say Jensen Huang, who you just had on the previous segment, believes that too. In fact, he thinks it's already here. So that's the first question. Will we get to that stage? And the second question is, can we ensure that we have an absolutely cast-iron guarantee of control? Here's my problem. This stuff is so complicated. It's one thing to be able to say to a restaurant, no, you're not meeting our compliance standards. is the government smart enough to be able to look at all of this and say, yes, you pass this test or no, you don't?

23:25How do we set that up? Who would be in charge of it? Who would staff it? I think that's a very important question. We have to work out the details as we have with the FDA and clinical trials. So it's important to remember the government does not conduct clinical trials, right? The government says we need evidence. Give us the evidence. The government sets the standards for those clinical trials, whether you have to, you know, have how long this has to be. Do you have to have people who are taking the dummy pills along the way or the dummy drills in the head or whatever may happen with that to do that?

24:02The government sets the standards for these things. Is the government capable of setting standards to say what's here and what's there? And how do we staff it with the right expertise? So I think we do need the government to bring in more expertise. We have within the National Institute for Standards and Technology, within NIST, we have what used to be the AI Safety Institute, was renamed Center for AI Security and Innovation, I think it is. And there are people there with good expertise, but probably we need more than that. But I think the real problem at the moment is that the technology is intrinsically unsafe.

24:45We're seeing from all the companies admissions that their systems are getting out of control. OpenAI nearly lost control of its entire compute infrastructure. It wasn't just Hugging Face, it was OpenAI itself. And OpenAI has many of the world's leading computer scientists. So if they can't control it, there isn't some other body of super genius humans somewhere else. You can always outwit the AI and keep it under control. So I think that the real question is, does this technology path that we've embarked on ever support the level of safety that we need? And an analogy that I've used is to a fictional past in which the airlines decided to breed larger and larger and larger birds to carry passengers.

25:43And they invested trillions of dollars in this technology. They succeeded in breeding birds with a wingspan of 250 feet that could carry hundreds of passengers. Incredible fuel efficiency. Two sheep and off you go across the Atlantic. Very, very quiet. The only problem was that the birds keep eating the passengers or dropping them in the ocean. And eventually we decide that this technology path, despite these wonderful advantages of fuel efficiency and low noise, could never support the level of safety that we would require. And I'm afraid that this is the case for the technology path that we're on now.