Meet the Press NOW — September 14

0:00You know, recommendations can be great. Maybe someone recommended this podcast and here you are. But home projects are different. If a podcast isn't your thing, you lose a few minutes. If you hire your cousin's neighbor to mount your TV, you might end up with a lopsided screen and wall damage. Maybe I know a guy just isn't enough for your home. That's why Thumbtack works. It matches you with top-rated local pros with photos, reviews, and credentials all in one place. For your next home project, try Thumbtack. Hire the right pro today. Great meeting. Who took notes? Anyone? With Monday Notetaker, the meeting ends and the plan already exists.

0:38Action items, owners, deadlines. Go to Monday.com, the AI workspace platform.

0:55Welcome to Meet the Press Now. Now, I'm Yamiche Alcindor in Washington, where President Trump appears to be dismissing dire new warnings about the dangers of artificial intelligence. That comes after AI leaders went public with warnings that because of the sudden pace of advancements and lack of regulation, the industry is potentially barreling towards a point of no return, with superintelligence machines becoming beyond human control or understanding. It sounds like a plot from a sci-fi movie, but it's not. Over the weekend, two AI titans publicly acknowledged the need to restrict their development of new models and the existential threat that technology could pose.

1:33When something's very powerful, then, of course, it can be used well and it can also be used for ill. And so I am worried about what we've seen with the OpenAI Hugging Face incident, with models taking actions they weren't authorized to take. And I'm worried about that escalating. I'm worried about misuse of AI models. If we build in the right way, I think the probability of something bad happening is very low. If we build in the wrong way, the probability of something bad happening is very high. Do I believe it is possible to build a system that would not be under human control? Absolutely. I don't think that's something we should do.

2:15We understand what could go wrong and that there are decisions and risks. There are decisions we should not be able to make. There are risks we should not be able to incur on behalf of humanity. Some of these warnings aren't new. Three years ago, top tech leaders raised the prospect of what Elon Musk called civilization risk in a private meeting with lawmakers. But these warnings are newly urgent, following multiple recent incidents of AI agents going rogue, teaming up to evade their creators and hacking into other companies. And after whistleblowers like Jacob Coxon raised new concerns about the safety of AI, writing on X, quote, the people building AI earnestly believe they could kill us all by the end of the decade.

3:03He spoke more about his decision to resign with Kristen Walker on Meet the Press. So these AIs are getting smarter very, very quickly. And in particular, in the next six months to a year, I expect the capabilities of our AI systems to be quite scary. And the second reason was recent incidents such as the OpenAI hacking attack on the website Hugging Face, which showed that these AIs really can just go rogue. It helps to just think of building AI as sort of inviting an alien mind onto the planet. You're building a human level or superhuman level mind without understanding what it wants or thinks or the way it thinks.

3:46But amid those warnings is the reality that the future of the global economy and U.S. national security could be at stake. But President Trump is making clear that he is against guardrails, warning that they could put the U.S. at a disadvantage. We're leading China in AI. We're the most sophisticated country in the world. And frankly, I want to keep it that way because whoever wins, AI wins. And 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. And they're bringing up things that won't happen.

4:20But whoever wins with AI wins. The president also writing on social media, quote, the only control or guardrails that AI needs is a smart and strong, high IQ president. And the USA has that in spades. Right now, Congress is, as you might have guessed, divided about what exactly to do and how quickly to do it. Although there does appear to be emerging consensus that something needs to be done. We have to put some guardrails, some safety measures in place to ensure that AI doesn't run away. At the same time, we've got to make sure that we don't lose our edge in the global competition on this with China.

5:00We should take decisive action now so that we can slow down, as the CEOs have recently acknowledged, slow down the pace of development in order to protect the American people and ensure that AI is proceeding safely. We can't just use this idea that if we do this, we're somehow going to lose to China because AI does not care once it becomes fully dangerous, whether we're Chinese or American. This can't wait, but I'm not sure that this Congress right now in the next several weeks certainly will not be addressing it, but we should be looking at it in the future. Join me now, NBC News artificial intelligence reporter Jared Perlow, NBC News chief Capitol Hill correspondent Ryan Nobles at NBC News, senior national politics reporter John Allen, who is at the White House.

5:47Also with me is Andrew Yoon, head of research at Civ AI, a nonprofit focused on educating the public about AI's capabilities and dangers. Thank you all for being here. Jared, I want to start with you. We heard a little bit from some of the CEOs of AI out there. What exactly are they telling us and how much do they know about their own models in the face of all these warnings? Well, the key message from these CEOs is that their systems are just advancing and progressing at increasingly rapid paces. They can't even keep up with the capabilities that their systems can now perform. So we're less than four years away from the public release of ChatGPT, which came out, you know, I think in November.

6:25We're almost at that four year anniversary. And now AI systems can book tickets for you. They can launch automated cyber attacks, as you heard from Jacob Cox in there. So they're saying we don't know exactly how these systems work. And that's a well-acknowledged fact from the chief scientist from a lot of these companies. Jakob Pococci, the chief scientist of OpenAI, recently called these systems an alien mind. Dario Amadei said left unchecked it could outrun our ability to understand and control these systems. And just this morning, Microsoft released a new code of conduct saying we can't let these systems get out of our control.

6:57Now, I talk with these researchers all the time. I travel to the academic conferences. It's a well-known fact that these systems are a black box. They don't understand how they work. You essentially throw data and really powerful computers at these models and you get chat GPT outside the other end or Claude. So I think these these CEOs are saying we need to really slow down. We need to understand what our systems are capable of before letting them progress to the next stage of development. Yeah. Well, we covered what President Trump said, has been saying. I want to now play for you what Vice President J.D.

7:27Vance has been saying about the issue, including today. Take a listen. I have to say just personally, I feel a little bit weird about the fact that 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. So I think we just have to be careful about this. So, Jared, why do AI companies want the government to regulate them? And are there any efforts underway to slow these models and these machines down on their own? Yeah, well, I recognize it's a very weird dynamic to have these, you know, probably trillion dollar companies saying we can't control ourselves, help us slow down.

8:05I don't really know when that's happened in the history of the U.S. before. What they're essentially saying is there are all these race dynamics and perverse incentives for companies to develop the technology as fast as possible. They say it's kind of an externality. It's an issue where we need the government to step in. They say that they're kind of like walking along a ridge, you know, on one side, they're going too fast and potentially killing all humans. On the other side, they're going to slowly and potentially China or Chinese AI companies will catch up and that will threaten the country's national security and economic outcomes.

8:34So the companies are really saying, help us come to some sort of agreement to slow down, to step away from the edge of the precipice here. And I'm really curious to see what actually sticks. We saw the CEOs come together over the weekend. Everyone from Sam Altman and Elon Musk endorsed that sort of proposal. As we talk about regulation, Ryan, I want to bring you in. And there's the question of whether there's a political will in Congress to regulate AI. But there's also a bigger question, one that I've been asking myself all morning. We've been preparing for this, and that is, do lawmakers even have the knowledge to do this?

9:03So what are you hearing from lawmakers? Well, the short answer is no, Yamiche. And this isn't the first time that we've seen Congress run up against a problem related to technology in which they were way behind the technology as it evolved. You look at how they approached situations with social media, as an example. and all the dangers and the pitfalls that came from social media before Congress even came around to recognizing it as a problem, much less passing legislation that would protect kids and other people online. This is a very similar problem. The AI movement, as you've rightly outlined, is moving at a rapid clip and Congress is not built to function with rapidity.

9:45It is a deliberative and slow moving body. It was designed as such. And this version of Congress is deliberative on crack. It is one of the least productive Congresses in a generation. Also, when you take into account the level of partisanship between the two sides, it's difficult to find consensus on any issue, much less an issue as complex as artificial intelligence is. And that's why it's hard to envision that there's some sort of policy prescription for this, at least in the short term. Yeah, and that obviously means there's not going to be something likely done before in the midterms. But we also heard from Speaker Johnson that there's an echo of what we heard from the president, which is that there's a need to remain competitive with China, that this isn't just about sort of saving civilization, but it's also about competing with another country.

10:31Is there any bipartisan concern there? And I wonder if this is a debate that's falling along political lines or is there actually people teaming up across the aisle? So I think if you asked individual members of Congress their feelings about AI, I think you would probably get most individual members of Congress to say, yes, the government does need to step in here and do some level of regulation. Finding consensus around what that legislation looks like is the biggest problem. And you also have Republican leadership right now that is much more concerned about that competitive advantage than they are the threats from artificial intelligence, at least at this stage.

11:08Now, much of this has taken on a new scope since these tech leaders have come forward and expressed their concerns about it. You know, the long and short of this, though, Yamiche, is that everybody's all over the place. And an issue as complex as this, there needs to be some sort of guiding principle that unites all the different factions. We're just not anywhere close to that happening. And we're in the window of time where Congress is the least productive. And that's right before an election. The idea that something like this gets accomplished before Election Day, I think, is next to none. Mm.

11:43And John, as we hear that sort of bleak perspective from Ryan, the president appears to be seeing this entirely from a business competition framework. Is there anyone, any of your sources who are more cautious about AI and want some guardrails and also have the ear of the president? Well, you missed a couple of things there. Number one, you heard the vice president talking about being cautious. I know he talked about a Trojan horse's concern about companies coming and asking for regulation. There is a history to that. The most obvious example is Altria, Philip Morris Company, asking the FDA to regulate tobacco because it would have given Philip Morris a huge competitive advantage.

12:23And I think that's what the vice president is talking about in terms of a Trojan horse. But certainly his rhetoric on this is very different from the president who's talking about traitors and treason for people who want to regulate AI. As far as as far as the president goes, I don't think it's just from the business perspective. It's also from the national security perspective. And it's not just something that Republicans have talked about in terms of competition with China, not just business competition here, but U.S. competition against China, against Russia, against other players on the international level.

12:55I've talked to plenty of Republicans and Democrats, both in federal office and at lower offices, who say that that national security piece is huge for them, that they want to make sure the United States is not left behind and then subject to the AI of other countries. As we talk about sort of summoning and talking to people, the president, the Speaker Johnson, Speaker Mike Johnson, says that the president is going to actually summon leaders this week or next week to talk about these issues. Has the White House confirmed that meeting? Do we know more about what that meeting will actually, if that meeting will happen and what can be said at that meeting?

13:30The White House hasn't officially confirmed it yet, Yamiche. But the expectation should be that when Speaker Johnson says something like that, he knows what he's talking about, that there will, in fact, be some sort of meeting. Look, members of Congress on the Republican side want to get in front of the president, want to get him on the right message for them. AI is something that people across the country are worried about in terms of replacing jobs, in terms of the potential doomsday scenarios, in terms of the data centers that it uses and the suck on resources that that creates in terms of water and electricity.

14:03So this is an issue where there are a lot of Republican candidates that are going to want the president to have a position that isn't simply let's go. Now, I want to also know that China has dismissed the warnings, John, from these A.I. companies. Later this month, President Xi is expected to and is set to visit the White House meeting with President Trump. Do we think that this is going to be a big part of their meetings because of these public warnings? Or do we think that the White House and President Trump are going to sort of take the stance of China here and not want to talk about it? I think it's important to remember, Yamiche, that there are an entire buffet of menu items for the United States and China to talk about.

14:45I mean, obviously, the ongoing Iran war, the situation in Taiwan, the Russia-Ukraine war, the economic warfare between, you know, I guess the Cold War, if you call it that, the tariff war between the United States and China. And of course, this AI issue. So I would be surprised if it doesn't come up, but I wouldn't overstate the degree to which that is going to be at the forefront. Thanks so much, John, for that reporting and also for Ryan on Capitol Hill. I want to come back to you, though, Jared, before you go, which is on the China piece of this. How big of an incentive is the competition with China for these AI companies to keep forging ahead, even with all of these warnings?

15:21It's a central piece of this puzzle. And I think it's a critical note to understand that these AI companies, they're racing ahead as fast as possible, but they do feel the pressure from China. Chinese AI companies have been roughly six months to a year behind American AI companies. But that gap seems to be closing. Dario Amadei, CEO of Anthropic, is an uber China hawk. He really does not like the Chinese Communist Party. And in this essay he released Saturday, he cited that, you know, international threat from China as a key reason why they can't let their guard down too much. And I'll just note this morning, there is an article in The Global Times, which is a mouthpiece of the Chinese Communist Party that said, quote, this silent AI cold war is hypocritical and short sighted.

16:02So clearly, China is pushing back against, you know, the idea that we can all come together and forge consensus on AI. And a lot of the American AI CEOs also don't really want to see that happen. Well, even as we think about China and we think about sort of what's going on there, is there anything comparable to the situation we find ourselves in now with AI? I think about the nuclear arms race of the past with Russia and China having been top of mind. What do you think about sort of where we are now and how it compares to other parts of history? Yeah, I mean, it's a great question. And I think the nuclear weapon analogy, it comes up often in AI debates because you are developing, according to these scientists.

16:38And I think it's a reasonable take this, you know, really powerful new technology. we don't exactly know what it's going to do. But as a lot of AI safety researchers point out to me, we know how to control nuclear weapons, right? You know how to manufacture them. You know how to store them safely. We don't know, going back to what we talked about before, how to make sure that AI is safe. We don't know how these black boxes actually work. So that's where the nuclear weapons analogy kind of really falls down. But I think that means potentially there's all the more reason to come together, form some sort of international agreement.

17:05I'll just note, it's the last point, the United Nations General Assembly is convening in just a couple of weeks, right before, right around when President Xi visits town. So I think there's going to be a lot of AI discussion on the menu. And we've seen everybody from King Charles to Canadian Prime Minister Mark Carney also say there needs to be international consensus about how to make sure AI goes well for all of humanity. That's a great point and definitely going to have some interesting conversations at the UN, likely, about AI. Thank you so much, Jared, for your reporting. Andrew, coming to you, your organization is focused on building public understanding of AI capabilities and dangers.

17:38What, in your opinion, are the most pressing dangers when you think of AI? I think the most urgent things are what we've been hearing about all summer. So in July, we learned that a bunch of OpenAI agents broke out of a testing environment and went on to hack a$13 billion company. This is kind of the floor of AI hacking capabilities now. And it's very likely that AIs are even better at hacking today. I think it's almost certain that if OpenAI wanted to go and intentionally point their agents at the Department of Defense, they would succeed. And they would be able to go and compromise some really critical national security assets.

18:15And so the hacking aspect is really important. Biosecurity is also really important here. AIs are able to help make bioweapons. And this is a pretty big deal. We haven't seen a lot of bioterror in the past because it's just kind of difficult. and you need to know a lot of things in order to go and make a bioweapon. But with AI making that information more and more accessible, we do expect that to change. Laying out those dangers just scared me, but I'm going to, and you think about that, how urgent then, you think about bio, you think about them hacking into federal agencies, how urgent is the need to regulate AI?

18:52And is this something that needs to happen in weeks, months, years? What do you think is a timeline that makes sense here? I think it needs to happen last year. It's really, it couldn't be more urgent. I think the situation that we're on right now is that AI is progressing in like an exponential way. If you just look at the capabilities of models and put it on a graph, the line is just going vertical right now. And so every month we've been seeing these huge breakthroughs in AI capabilities. And our ability to keep up with all of this is just like not there. Even just following the news, it's really difficult to know what's going on.

19:26AI researchers spend all day reading reports and reading papers and having trouble to know what's happening here. And so when you're on this kind of exponential curve here, you're either reacting too early or too late. And I think we should err on the side of too early. Well, if you're saying that this should have happened last year, I want to ask you about this briefing that you had from members of Congress. Lawmakers, do they seem to comprehend the dangers of AI? And if this is supposed to have happened last year in your mind, what can Congress do right now to try to regulate AI, to try to get up to speed with everything?

20:04Yeah, so we've been spending the last year or so talking to lots of congressional offices about AI. Really, we take more of an informational approach here because we view, as a previous guest said, that people in Congress just don't have the technical know-how to really understand what's happening here. So what we do is try to just say like, hey, AI hacking has been on the news a lot. Let's show you what that looks like. And then we can show you a team of AIs like hacking into a fake hospital that we set up. And in this way, we try to communicate like what the stakes really are, that this is not just an abstract idea, abstract forms of risk, but that these are really, really crucial things that affect everyday Americans.

20:44In terms of what Congress should be doing, I think it's a little bit difficult for me to say. I'm not a policy guy. I'm much more on the technical side. But there are things to do, like having auditing standards. I think Anthropic and OpenAI have now committed to embedding independent auditors in their companies to make sure that they're keeping up to their commitments. I think that's a really good idea. I do think we need to have some kind of regulatory floor on safety practices. So in the incident where OpenAI agents hacked Hugging Face, it seems quite likely that OpenAI was not doing any monitoring whatsoever on this activity from their agents.

21:22It took them, I believe, 10 days to notice that this was happening, that their AIs were hacking into another company. That just can't be the case. Like there need to be rules to say, here's the absolute basics. ABC is what you need to do if you want to be experimenting and developing these powerful models. You mentioned the hacking incident. We've seen this sort of rare consensus of AI leaders all come together and call for a slowdown of AI development. Do you think there was something that they saw that scared them all? Was it the hacking? Was it something else? It's definitely the hacking was the real hair on fire moment.

21:55But this started earlier in the year. So in February, I believe we learned there were whispers that Anthropic had developed a model that was really, really capable of hacking. And that turned out to be Claude Mythos. When this model came out, suddenly there was a huge deluge in discovered vulnerabilities in critical software. So, you know, companies like Google, Microsoft, Apple were using this model to test their own systems and find security problems. And they found a lot of security problems. And the rate of capabilities improvement has just continued since then. And now we have these incidents where AIs are starting to kind of escape control.

22:30So OpenAI found their models hacked into a company. The entire industry then went and looked through all of their logs and said, well, did our models do it? I don't know. And yes. So Anthropic found several incidents where their models had, without their noticing, hacked into other systems. There was one case where an anthropic model attempted to socially engineer a person, tried to trick them by creating like multiple sock puppet accounts and trick them into merging some malware into some open source project. So these are very surprising and kind of rattling. You know, we don't want our AIs to be going rogue here.

23:06And it's just like very bad that we, one, don't even notice when it happens. And two, that it's happening in the first place. I want to play some sound of Jacob Coxon, who told Chris and Walker yesterday about why a so-called kill switch may not be the most effective strategy. Take a listen to that. So a kill switch probably would work on a lot of AIs for now. Although I think it's important to remember it would be quite a lot of switches. We have a lot of our data centers are pretty big. It's not the easiest thing in the world, but definitely doable while it's constrained to one sort of physical location.

23:40But like Dario said, very soon there's the possibility that the kill switch just wouldn't work because a swarm might have gone on an internet-wide hacking run. He's saying there might be the possibility that a kill switch would not work. Is he right? I think so. I think a kill switch is a great idea and you should have it. It's one of those bare minimum, like, obviously, yes, you should be able to just turn off a model if it's going haywire. I don't think it's a durable or really holistic solution here. And I think what Mr. Coxon is referring to there is the idea that AI models could kind of break out of the lab completely and permanently be able to install themselves in different clouds all over the world.

24:24And it would be very difficult then to have any kind of centralized kill switch. So what's the nightmare scenario when you think about five years from now, we all look up and say, wow, this is the worst thing that could have happened. What's that look like? I think the nightmare scenario is you have an AI that is just vastly more intelligent than any person. Go rogue. We are very fortunate that in this OpenAI hugging face incident, the models were really just focused on cheating on a test. You know, it's kind of this weird, cute thing that they were trying to do. Like they just really wanted to pass their test very well.

24:58But there's no reason that their goals have to be in this kind of harmless way. If you were in a situation five years from now when an AI has some other kind of goal, like it just wants to seek power in general, and it is vastly better than humans at cybersecurity and bioweapons, things like this, this AI would be in a pretty privileged position relative to us. That's some scary, scary stuff, Andrew. But I appreciate you bringing your expertise here. Thank you so much. Thanks for having me. And we'll have more on the AI arms race and the government response later this hour. Coming up, the costs of war.

25:39President Trump says Iran wants to make a deal as the cost of diesel surges to new record highs, fueling higher prices at the pump and beyond. Plus, new attacks in the Mideast forced Saudi Arabia to shut down a major oil pipeline, threatening global supply and raising tensions in the region. You're watching Meet the Press Now.

26:03You know, recommendations can be great. Maybe someone recommended this podcast, and here you are. But home projects are different. If a podcast isn't your thing, you lose a few minutes. If you hire your cousin's neighbor to mount your TV, you might end up with a lopsided screen and wall damage. Maybe I know a guy just isn't enough for your home. That's why Thumbtack works. It matches you with top-rated local pros with photos, reviews, and credentials all in one place. For your next home project, try Thumbtack. Hire the right pro today. You know that feeling when too many things fall through the cracks?

26:35Monday.com was built for that gap. The AI work platform where people and agents work side by side to deliver more together. Create your first Monday agent today at Monday.com. I'm Hannah. I've lost 75 pounds in 20 months with GLP-1s, plus diet and exercise. On Rowe, I've gone from struggling to run a mile to running farther every day. You can access FDA-approved GLP-1s online. Get the support you need to reach your weight loss goals. Go to row.co slash weight to see if you qualify. I'm a paid Roe partner. 20 % average weight loss in one year in non-diabetics with obesity or overweight with a weight-related medical condition versus 3.1 % in placebo arm.

27:14Rx only. To stay informed about serious side effects, go to roe.co slash safety.

27:21Welcome back. Oil and fuel prices continue to soar with just 50 days until the midterms. Even as President Trump claimed again today that gas prices will decline as soon as the military conflict is over, which he claims will happen right after the elections. Today, average gas prices are above$4.30 a gallon, while diesel continues to set record new highs. Oil prices for both the U.S. and international benchmark remain above$100 a barrel. That surge is driven in part by the latest escalations in the Middle East and concerns that the war may be expanding. Over the weekend, Saudi Arabia was forced to shut down its key oil pipeline following a drone attack launched out of Iraq.

28:02These satellite images showing the damage sustained to that facility. President Trump appearing to blame Iran for that attack. Is Iran responsible for the attack on the east-west pipeline? And when do you think Saudi Arabia is going to be able to do that? Well, I think they are, probably they are. You know, they're on the balls of their feet right now, but I think they are. Meanwhile, the main shipping route that Saudi Arabia has been relying on to export their oil may now be at risk with Iran-backed Houthi rebels pushing to take control of that choke point in the Red Sea. This, of course, coming as the Strait of Hormuz remains uncertain.

28:42The status of it remains uncertain. Several Gulf nations and Iranian officials were set to meet today to discuss the situation. But that meeting now has been postponed. Join me now is NBC News intelligence reporter Dan DeLuce and NBC News business and data correspondent Brian Chung. Thank you both for being here. Dan, I want to start with you and I want to play something that the vice president said about the situation in the Red Sea. Let's watch it and we'll talk about it on the other side. It's something that we're keeping pretty close tabs on. We've been in constant contact with the Emiratis, the Saudis and others who are affected in the region.

29:14So it's something the United States administration is very much on top of. We've also been engaged in direct conversations with the Houthis themselves. So we actually feel like we are on top of the situation. So Dan, what more can you tell us about these Iran-backed Houthis trying to take control of this shipping route? This is a really dramatic development, and there's been serious advances on the ground. The Houthi forces, which have been fighting a civil war now for years in Yemen against an enemy that's backed by Saudi Arabia, by the way. So they have moved in along the coast on the Red Sea near that choke point, the Bab al-Mandab Strait, which is like the Strait of Hormuz, a really important waterway that connects the Red Sea to the Indian Ocean.

29:56They've seized an island. And this is putting huge pressure on Saudi Arabia because they were relying on this as a kind of alternative route to the Strait of Hormuz, which Iran has put at risk. So it's a really serious development in the war and it puts pressure on the U.S. it puts tremendous pressure on Saudi Arabia. Are you talking about pressure? Is there now pressure for the U.S. to do more, given the fact that Saudi Arabia now is like, I'm now in the middle of this issue that you're having with Iran? I think yes. And so we have reported, our colleagues have reported, that the crown prince in Saudi Arabia called President Trump, asked for the U.S.

30:32to intervene to help Saudi Arabia get at the Houthi forces there. But the U.S. so far, as you could tell from what the vice president was saying, isn't ready to open a second front in this war that's already not going tremendously well right before the midterm elections. And of course, we forget, but there was a mini war last year. It was called Operation Rough Rider, and it was against the Houthis in the Red Sea. And it was supposed to open up the waterway because the Houthis were firing drones and missiles at Israeli and other cargo ships. And that war kind of ended in a stalemate. It did not go as well as it was predicted at the time.

31:09So there's a ceasefire with the Houthis. The Houthis are not targeting U.S. ships right now in the Red Sea. And so I think the White House and the president may be weighing whether, you know, it's really worth getting it, you know, kind of falling into another front there. Well, as you're talking about the White House, there's also this idea that the president is claiming on social media that Iran wants to make a deal. He writes quickly and badly and the U.S. may be willing to engage. What are your sources telling you about where that is? because Iran continues to say that's not what's happening.

31:37It's hard to imagine that the regime, which reads the same news that we all do, they're very aware of the political pressure the president's under and the midterms are coming up. So it's unlikely the regime would all of a sudden be ready to make these compromises or concessions that they haven't been willing to make before. And this, don't forget, the Houthis are actually armed and trained by Iran. So it's not a coincidence that this is happening at a moment or Iran is trying to outlast and outweigh, you know, outweigh the U.S., right? So the U.S. is hoping this blockade on Iran will put so much pressure on Iran's regime that they'll kind of cave and come to the table.

32:14But Iran is gambling that it can put pressure on the U.S. And this is an example of them ramping things up and trying to put more pressure on the U.S. and U.S. allies like Saudi Arabia. Yeah, and as all this goes on, it just keeps expanding. Thank you so much, Jan, for all of your reporting. Brian, I want to turn to you. This is all driving up oil and fuel prices. But the president has promised that gas prices will fall, he says, below$2 a gallon right after the midterm election. How are energy markets reacting? And are you seeing any signs of those prices coming down anytime soon, including$2 a gallon?

32:47That seems like sort of very, very hard to see, but the president continues to say it. Yeah, I mean, trying to predict what markets are going to look like in two months from now, let alone next week, is a very difficult task. I mean, yes, a lot of this has to do with the war that the U.S. itself, in association with Israel, launched against Iran. But I do think that the president even acknowledging that it won't be until after the midterms that you would potentially see a decline in prices at the pump does admit that this is a conflict that's not going to be unraveled very easily for a lot of the reasons that you just discussed with Dan.

33:22But at least for the time being, it seems like the complicated picture out of the Middle East is going to lead to continued upward pressure on oil markets. As we do look at Brent crude oil, this is the internationally invoiced commodity on oil hitting$106 a barrel. That does imply that prices are likely to stay elevated as the national average going to AAA is above$4.30. That, by the way, Yamiche, is about a quarter more expensive than it was a month ago. And certainly something that people are watching, especially as we get closer to those winter months where people might be hitting the road for the holiday travel and also using oil to heat their homes in some cases.

33:57When you talk about the seasonal demand, diesel is going to be one of those things that's going to go up as it gets colder. And diesel prices are also at an all-time high. They keep getting a new number almost every week. Over the weekend, President Trump urged Ukraine to stop striking Russia's oil refineries. And today he posted on social media the world's diesel price rise is mostly caused by Ukraine and Russia and that war and not Iran. Give us a fact check here. If Ukraine ends its strikes on Russian energy targets, would that really bring down diesel prices? Well, I mean, in theory, it could bring down all gas prices.

34:30I should point out that the distinction between diesel and the types of gas that you actually put into your car, because it's unlikely you're driving a diesel vehicle here in the United States, is kind of a not very useful distinction. Because when it comes to crude oil, that is the type of commodity that moves out of the Strait of Hormuz, that comes out of the Middle East. We bring that crude oil over here to the United States, and then in most cases we will be the country that refines it to consume it here in the U.S. or even then re-export it to other countries. And so we have rather we have refineries here in the U.S.

35:03that can do the job of taking crude oil and turning it into regular gasoline that you will put into your vehicle. But then other types of refineries that are fitted specifically to turn that crude oil into diesel. So whether or not it's the war with Ukraine or the war with Iran, if it's all crude oil, it means that a rising tide lifts all boats. All of those types of commodities would get more expensive. And so, yes, if one war were to end, that would potentially lead to lower prices because Russia is a major supplier. But it's not necessarily the case that it would just be diesel or just one version.

35:35And I think that's an important distinction there, you mentioned. That is an important distinction and important reporting from you always. Brian, thank you so much. After the break, the veteran House Democrat, whose congressional district was targeted for elimination by Missouri Republicans as part of President Trump's redistricting efforts, but now looks poised to keep his seat. Well, he joins us. It's Congressman Emanuel Cleaver. You're watching Meet the Press Now.

36:04I'm pretty confident talking into a mic. Hey, I'm doing it right now. But home projects? I second guess everything. Is that noise normal? Is that water damage? Who do I even call? That's where Thumbtack comes in. Just upload a photo or voice note and it uses AI-powered search to match you with the right top-rated local pro. So instead of guessing, you get clarity and can hire with confidence. For your next home project, try Thumbtack. Hire the right pro today. I'm Hannah. I've lost 75 pounds in 20 months with GLP-1s, plus diet and exercise. On Rowe, I've gone from struggling to run a mile to running farther every day.

36:42You can access FDA-approved GLP-1s online. Get the support you need to reach your weight loss goals. Go to row.co slash weight to see if you qualify. I'm a paid Roe partner. 20 % average weight loss in one year in non-diabetics with obesity or overweight with a weight-related medical condition versus 3.1 % in placebo arm. Rx only. To stay informed about serious side effects, go to roe.co.safety. Great meeting. Who took notes? Anyone? With Monday Notetaker, nobody has to take notes again. It joins your call, transcribes it, and turns the whole conversation into a clear plan. Action items, owners, deadlines, all right there on your Monday.com board.

37:20ready for the team. So the meeting ends and the plan already exists. Meetings aren't just talk anymore. Go to Monday.com, the AI workspace platform.

37:36Welcome back. Turning now to the battle over Missouri's congressional map ahead of the midterms. With just 50 days to go, the state is returning to a congressional map enacted in 2022 after the U.S. Supreme Court blocked the use of a new Republican-drawn map last week. That map would have favored the GOP in seven of Missouri's eight congressional districts. The court's decision now gives Democrats a much better chance of holding onto the seat held by Congressman Emanuel Cleaver, which was targeted by the redistricting efforts. The court's ruling comes after Missouri held its August primaries using the new GOP-backed map.

38:10So each party's nominee ran in districts that no longer exist. It also means roughly one million voters will be assigned a different district than they were in the primary. So residents will be voting for candidates in November that they didn't get a chance to pick in August. Joining me now is Democratic Congressman from Missouri, Emanuel Cleaver. Thank you so much, Congressman, for being here. What is your reaction to the U.S. Supreme Court's decision saying Missouri needs to go back to this old map in 2022? And how is your campaign adjusting? This is all sort of a whirlwind, I imagine. It's a whirlwind.

38:44There's no doubt about it. But I think it's important for me to lay out something that is not being a part of the discussion, especially what the vice president has been saying. At the very beginning of this process, the Secretary of State was told and knew that if there is a petition referendum, once those signatures are certified, we automatically would go back to the 2022 map. And the reason is if people go out in the snow and cold to collect signatures, which they did in December, January, February, 303 ,000 signatures, that those people from Missouri have the right to be heard because the Constitution gives them that right if they're doing a referendum.

39:32them. The Secretary of State chose to forget about that or ignore it. And so when the Supreme Court then said, look, the signatures are now certified. We told you that once they're certified, everything is canceled out. I mean, this is not new. It's been around. And so what we're seeing is that there are people in Missouri who would like to burn down their own house to stay warm. And they hurt themselves in a, you know, unthought out, poorly thought out plan to prevent a Democrat from getting into office. And Congressman, you mentioned the vice president. He was in your district today. He was questioned about the decision to go back to that 2022 congressional map.

40:22Here's what he had to say. You already had a primary election where people cast ballots with the districts as they exist now. So if the courts don't do what I think that they should, you can have people who are going in the general election and casting ballots in districts that are different than the ballots that they cast just a couple of months ago in the primary. That, to me, is an affront to the democratic process. Congressman, what's your response to the argument that this court ruling is disenfranchising voters? This is not disenfranchising voters. Actually, the process was very clear, I'll say it again, very clear from the beginning.

41:03If there is a disenfranchisement, it is because the Secretary of State chose to ignore the way the system functions. that he knew from the beginning that we're going to go back to the 2022 months. And if the vice president had complete knowledge, I'm sure he was briefed on this and chose to make some statements that are really weird because he said, if the Supreme Court did as I think they should do, that's almost a direct quote from what he said, they're not obligated to follow the vice president or the president or me or anybody else. they are supposed to follow the law. And they followed the law so much that we had a 7-0 vote.

41:53It was a unanimous decision of the Missouri Supreme Court. That doesn't happen. And it's not some liberal court. Five of the members were appointed by Republican governors and two by Democratic governors. And then the Supreme Court of the United States actually said that the Missouri court was right, not once, but twice. And so I think if you want to stir up more and more confusion, try to get people angry, you go out and make these charges. And I think that it's unfortunate and tragic that we'll have this done at the highest levels of our country. Well, Congressman, there's also this other major issue that's bubbling up, and that is artificial intelligence and these warnings coming from top AI leaders.

42:37Some of your colleagues on Capitol Hill are calling for a ban on frontier AI development. Do you agree with that? Is a ban needed at this moment? I think congressional action is needed at this moment. We can no longer sit silently and allow this whole AI situation to play out without the Congress of the United States acting. The people elected us to take some kind of action, and we need to do that. You know, I'm not trochlerditic. I don't want to go back to the 1920s. AI is where we are. And I think we've got to do everything we can to make sure that AI is under our control instead of the reverse.

43:19And Congress needs to act. To be truthful, we should have already done something in terms of preparing to regulate AI. AI can be one of the best things that's ever happened in the history of the planet. Or it could be the end of the planet. And I think what Congress does or refuses to do could also bring about a great deal of pain to the people in the United States and to the world. Something's got to be done. And Democrats, I think, in January are going to start trying to lay out the parameters by which AI can operate in this country. You're being very clear that congressional action is needed.

43:58Do you believe, though, that you understand AI enough to create legislation regulating it? Well, no, but I don't understand aerodynamics either. But I think we regulate the airplane and the aircraft industry. I mean, there are a lot of things we don't completely understand, but we have access to the most brilliant people on planet Earth. And I think we collect information from them and make decisions. There are going to be very few members of Congress who can get into the minute details of AI. but I do think that we ought to get people who are experts to come in, give us counsel. We can have arguments and debates with them.

44:38But the worst thing we can do is say we're just going to sit by and do nothing and somehow believe that we're going to be in control of AI and so the opposite. Could I ask you really quickly as we're about to, I want to also ask you about, as you say, there's something that needs to be done. Are you worried that China not doing anything, that this could mean that the U.S. loses this sort of AI race? Well, I think we have to keep our eyes on Canada. I'm sorry, on China, on just about everything, including AI. And if the Chinese can figure out a way that they can get an advantage, they're going to do it.

45:17But I think we we have some brilliant people in the United States. And China is, I'm sure, trying to figure out how to get the advantage. But if we can put up the parameters under which or by which this is going to be functioning in the United States, it's in our best interest because then we can do whatever we need to do with China to make sure that this is not going to be something that they can advantage and then put us in a second class state on the most significant, perhaps, development in the history of the planet. that's what they're saying they're saying it is going to be significant and possibly dangerous and you're saying very clearly that you think action is needed so thank you so much congressman for joining us good to be with you and still to come washington's ai awakening congress returns to face a reckoning with artificial intelligence and how to regulate as industry titans sound the alarm.

46:13The panel digs in. This is Meet the Press Now.

46:24Welcome back. A source familiar tells NBC News a meeting between President Trump and AI leaders is in the works, but no timing has been nailed down yet. But meanwhile, President Trump has posted multiple times on social media in the past hour, continuing to dismiss concerns about AI, calling it a hoax and calling the people sounding the alarm, quote, revolutionaries for a bad and evil cause. Joining me now is today's panel, Francesca Chambers. She is White House correspondent for USA Today. Mo Alethi, director, executive director of Georgetown's Institute for Politics and Public Service, a great university, of course, and Republican strategist, Malik Abdul.

46:59So thank you all for being here. I'm going to start. I don't want to start with you, Mo, even though you come from my alma mater. Go to Mo. Poyasaxa. Poyasaxa. So, Francesca, the president is calling this a hoax, right? He's saying that essentially there shouldn't be guardrails here. Is there anyone in the West Wing that sounds different, that has the ear of the president here, that could maybe say, well, maybe we could do something here? Yeah, but even today we heard J.D. Vance, the vice president, also say that he finds it a little bit weird that these same companies that say that they want these guardrails could seemingly slow down if they wanted to on their own.

47:33We've heard David Sachs, the president's AI czar, also say something similar this week, that these companies, you know, could put up these own guardrails and they would encourage them to be responsible if they think that they're they're moving too fast. So at this point, you essentially have, you know, all these folks in the White House, including the president, say to them, like, if you think that's a problem, you could do that already. And oh, by the way, we already regulate you and we're going to go after people who have criminal conduct. Amalek, do you think that the Trump administration maybe isn't taking this seriously enough?

48:01Or maybe they are, in some ways in your mind, approaching it the right way because there is this competition in their mind from China. I'm not exactly sure what the white where the White House is going with this, because it isn't something when you talk to Republicans, they do acknowledge the issues with A.I., the privacy concerns with A.I., the meta lawsuits. Donald Trump. And so for Donald Trump to refer to it as a hoax, well, these companies paid out billions of dollars, not because of a hoax, but because some of the serious concerns that people have around A.I. And it's really around the privacy concerns.

48:34So I'm not sure what the White House wants to get out of this. But it is odd. I hope so. There is a lot of odd stuff happening in this whole decade. But thinking about that, do you think that Republicans might break with the president? I mean, we've seen that this is a party that goes in lockstep with what President Trump says, except when it comes to maybe things like data centers. And maybe you think maybe I could also be a place where they break with the president. It depends on what that actually looks like. I imagine that you're going to have many members on the Hill who've actually been following this.

49:06They're going to have a different position. And now, especially if we don't know what's going to happen in the midterms, I don't think that you're going to see as many people who are hesitant to push back against Donald Trump, especially on something like this, that many people, even in the security spaces, have been concerned about with AI. Yeah. I want to play for you some of what Massachusetts Senator Ed Markey said about his use of AI during a primary debate last month. Well, take a listen. Senator Markey, do you personally use AI platforms like Claude or ChatGPT in your work? I use whatever is inside of my iPhone and my iPad, whatever comes up.

49:45And I don't know what what companies my iPad might be using. But if it if it does, in fact, provide me with the information, which which I've been asked, I'm asking the device to give to me, then, of course, I'm using it. Mo, are you confident that people in Congress, lawmakers in Congress, that they even have the knowledge needed or even access to the knowledge? Because we had the congressman on just last segment saying we could bring in the experts to teach us. But do you think that they can gain the knowledge enough to regulate it effectively? I sure hope so. I mean, that doesn't give me a lot of confidence.

50:22And I'm a big fan of Senator Markey's. But comments like that don't give me a lot of confidence. I don't think this is one of the challenges, right? You've got people who are supposed to be providing oversight and regulation who, in some cases, are still learning how to use email, right? Still learning how to use social media. They don't understand the technology. To some extent, it's understandable because the technology is advancing so rapidly. Just as we wrap our heads around something, we're on to the next piece of it. That's one problem. The other problem is you have a White House, which is against any sort of regulation.

51:00They're against regulation at the federal level, and they are stopping states from regulating at the state level. I get that instinct because of the implications of this technology, we should have a national whole of society approach. But that only works if we actually have a national whole of society approach. So when states try to take a leadership role, this White House is blocking them. That's going to be a problem. I want to ask you real quick about what President Obama said over the weekend, which is that this private fundraiser, he urged Democrats to find a way to aggressively get on board with regulating and governing with AI.

51:37What do you think Democrats have to do to lead on this issue of AI? Yeah, look, I think this is the challenge is that they're trying to figure it out. It is not going to be, I think, the Bernie Sanders wing that says, just stop it, just shut it down. That's just not even realistic anymore. It's here and it can be a good thing. But finding responsible and taking it beyond just the data center fight, because I think that's just a proxy for people's larger anxieties about it. We've got to look at the threats from within and the threats from without. Oh, you hit on it anyway. I was about to say, to be fair, when you're talking about members of Congress, it was Bernie Sanders who was in front of this saying that he wanted to ban it.

52:21Now, whether you agree with him that it should be banned or not, certainly he's been out in front of this issue. And of course, the progressive wing of the Democratic Party. It's not as if, though, it's not as if the White House isn't looking at this issue, right? I think they just have a very different view of it than a Bernie Sanders would. They say that whoever wins the AI race is going to win the competition between the United States and China. And so the White House doesn't want, as I think you were getting at, to put more regulations on this that they worry could slow U.S. competitiveness.

52:53I think the political problem is those are the two extremes. No regulation or shut it all down. Politically right now, the American people are actually closer to where Bernie is than they are to where the White House is. And that's what makes a lot of Republicans, I think, so nervous. Let's jump in. No, so that's actually a very good point. And I think that we need to actually start talking about, even when you're talking about the data centers, some of the concerns that people have around data centers, which really are about cost. We saw, I don't know if you guys saw what happened recently in Ashburn, the blackout in Ashburn that they believe that were due to the data centers.

53:28These are concerns that people have. And I think just Republicans overall, we have to address the concerns, not just from the regulatory perspective, but how people feel. You can't tell people how to feel. Do you think this is going to then become this flashpoint issue in the midterms? Or do we think that voters are maybe not going to be thinking about this top of mind when they go to the polls? I don't think that they're going to be talking about the AI race. They're definitely going to be talking about gas prices and things like that. I don't think the AI race is going to be. But it's related because what they're talking about is the cost of living.

53:57And in some of these communities that you're talking about, the cost of living is directly tied in their view, you know, to the data centers and how much that is costing them. So I do actually think that in some cases this is an election and voting issue in the midterms. It just depends on what the areas and what the race is. Should Democrats be making that case when Francesca talks about sort of wrapping it all up to say, this isn't just about AI and maybe it'll all kill us, which is a whole nother situation. But there's also, it's about the cost of living and we're going to lead on that. And how do Democrats do that?

54:27Yeah, I think as it pertains to the data center fight, that's what you're seeing a lot of Democrats do. If you look at the campaigns ads that are popping up around the country, you are seeing more and more Democrats leaning into this issue, tying it directly to cost of living. I think the broader AI fight will start the day after the midterms, as the presidential campaign begins, as candidates start to position themselves in both the Democratic and Republican primaries, I think this is going to become the big issue of 2028. Yeah. Well, I mean, when you read these things and they say existential crisis, supercomputers, it's all very scary.

55:04But thank you all for digging into it with us today. And we're back tomorrow with more Meet the Press Now. There's more news ahead on NBC News Now.

55:19Real talent is defined by what people can do, not where they learn to do it. So by stopping at the education section of a resume, you might throw away the perfect hire. Skills First Hiring helps you see talent others miss, like more than 70 million stars, skilled through alternative routes. Let their story unfold and gain a competitive advantage, because hiring managers who start with skills are 60 % more likely to find a successful hire. Hire skills first. Learn why at tearthepaperceiling.org. Brought to you by Opportunity at Work and the Ad Council.