Return to Transcripts main page

CNN News Central

Russian Murder-For-Hire Plot?; A.I. Going Rogue?; Trump Threatens to Tear Down Kennedy Center; Diesel Disaster. Aired 1-1:30p ET

Aired September 17, 2026 - 13:00   ET

THIS IS A RUSH TRANSCRIPT. THIS COPY MAY NOT BE IN ITS FINAL FORM AND MAY BE UPDATED.

[13:00:00]

(COMMERCIAL BREAK)

[13:00:25]

BRIANNA KEILAR, CNN HOST: A diesel disaster for Americans, gas prices shooting up, and the fuel that companies use to move everything everywhere hits record highs. Even more worrying, many analysts fear the worst is yet to come.

Demolition man, President Trump seen with a poster that appeared to read "Kennedy Center Demolished" just hours before administration officials told the judge that the performing arts center needed to be closed.

BORIS SANCHEZ, CNN HOST: Plus, a murder-for-hire plot that targeted a Russian dissident on American soil. Officials say they have broken up an operation backed by the Kremlin.

We're following these major developing stories and many more all coming in right here to CNN NEWS CENTRAL.

KEILAR: Today, a growing problem for Americans on the road and Republicans on the campaign trail, the average cost for a gallon of gas surging 7 cents overnight to $4.44. That's almost $1.50 more a gallon than before President Trump began the war with Iran.

It all adds up. You now have to pay about $22 more to fill up compared to just six months ago, and the sticker shock isn't only at the pump. Surging diesel means the cost of moving things is increasing. Every package, every delivery, every grocery item just got more expensive.

CNN's David Goldman is tracking this trend for us.

So, David, what's the latest here?

DAVID GOLDMAN, CNN BUSINESS SENIOR REPORTER: Yes, well, you're right.

I mean, this is a record, $6.40 a gallon. And think about what it costs just to move stuff across the country. Like, if you were -- we don't grow oranges here in New York, say, and you want to move those from Miami, that's 1,300 miles away. That's like $1,200 just to push a truck all that distance because of the price of diesel.

That's twice as much as it cost before the war. This isn't helping either, $4.44. We're only 10 cents away now from the high that we saw during the war, and about 60 cents away from the all-time high. That isn't given folks any comfort as well.

Our friends at Brown University have a tracker going. It is almost $109 billion the cost for Americans just for gas and diesel during the war, and that really adds up. That's $820 per U.S. household just from this.

Now, President Trump and Mike Johnson said the other day, well, when the war ends, then the price will come down. And that is probably true, because we saw that happen in June when the Strait of Hormuz reopened. But it only reopened for three weeks.

And, since then, this war has gotten significantly more complicated, and it's not just all that easy to just open up the Strait of Hormuz, clearly. This is a problem that's just mounting for American consumers, and this is a big reason why the Fed had to raise rates yesterday.

KEILAR: And, David, after yesterday's rate hike, we have seen stocks recovering a bit of ground here. What does it tell you?

GOLDMAN: Yes, well, I mean, I think that this is a good sign. Clearly, they were down a lot yesterday because the Fed raised rates. Doesn't like that. And it's signaled that it would have to do it again. But I'm going to suggest that it has less to do with the Fed and more to do with this.

There is some good news happening today. The Saudis said that it -- they might be able to get some of the oil flowing again through a pipeline that the Houthis attacked that's lowering oil prices. That is a breath of fresh air, but it's only about half of the oil that had been canceled in that pipeline because it had blown up.

Until this is really solved, that means ending the war, well, we're going to see oil prices in a very, very high range for a while.

KEILAR: Yes.

All right, David, thank you so much for that. Really appreciate it -- Boris.

SANCHEZ: In his fight to have the Kennedy Center bear his name, President Donald Trump is digging in. A new court filing reveals that the administration has temporarily closed the performing arts center to -- quote -- "assess existing safety risks."

It says the closure is going to last seven days, but it does add that that time frame could be extended. Notably, the president was photographed last night through a window on Air Force One looking at a poster that appeared to include the words "Kennedy Center Demolished."

Now, CNN has asked the White House to clarify what Trump was reviewing. No direct answer on that yet. Important to point out that Congress has already approved $257 million to renovate the memorial for president John F. Kennedy, but President Trump says it is going to take more.

[13:05:10]

(BEGIN VIDEO CLIP)

DONALD TRUMP, PRESIDENT OF THE UNITED STATES: For the Trump administration, for me to get involved and to take that on long term and to carry it or to raise money to carry it, I think that the Trump administration should certainly have recognition, because, frankly, if we don't do that, it's going to close. It'll end up being ripped down.

It's in very, very bad shape.

(END VIDEO CLIP)

SANCHEZ: CNN's Sunlen Serfaty is with us now.

So, Sunlen, where do things stand in the legal fight?

SUNLEN SERFATY, CNN WASHINGTON CORRESPONDENT: Well, things, Boris, are, frankly, very legally and politically messy right now over the fate of this building. And, frankly, things have escalated rather quickly in the last 24 hours, not the least of which, because you have the president out there warning that the building potentially could be ripped down.

We have him, as you noted, looking at that picture that allegedly said Kennedy Center demolished. Now, here's where we stand. This morning, we saw the lawyers in the administration file an update, noting that the building, they say, will only be closed for seven days. It's temporary. They say it's due to the safety concerns.

And in that filing, we had the executive director, Matt Floca, in a sworn statement saying -- quote -- "As the official responsible for the executive operations of the building, I determined that continued life safety hazards to individuals who might access certain public portions of the building have created untenable risks to the health and safety of those individuals."

But, notably, that if they -- officially close, they violate the court order, Congresswoman Beatty's lawsuit that says while this works its way through the courts that it needs to remain open. So, of course, we see a response from Beatty and her legal team saying -- quote -- "A temporary closure now is a recipe for defendants extending their shutdown forever and closing the Kennedy Center for the next few years, the very thing this court said they could not do without its permission."

And they, Boris, here are really wanting the judge to make crystal clear and stand in the way of this building demolished by those threats -- with threats from the present themselves. And so we wait to hear what Judge Cooper in this case will do, yes. SANCHEZ: Meantime, there's a coalition of preservation groups trying to challenge these plans, and they're asking a judge to put the center under receivership. Walk us through what that would entail.

SERFATY: Yes, this is interesting.

This is a separate, entirely separate lawsuit that they have already brought suit against the Kennedy Center officials here. But, today, they really sounded the warning alarms, saying, look, if it needs to get to this point, we will request for the judge to step in, appoint someone new to look over the Kennedy Center.

They say, if the current leadership of the Kennedy Center is unwilling or unable to exercise their duty, they call on the court potentially to consider appointing a receiver or replacing that leadership. So that would be a very dramatic step, so indicating potentially this could be one more lever to pull in this very feisty legal battle that's playing out.

SANCHEZ: Sunlen, Sunlen Serfaty, thank you so much for the update.

Still to come: A.I. going rogue again. OpenAI now reporting several more instances of its models being deceptive, doing things without permission. We have the details.

Plus: Russian spies and a murder-for-hire plot, the Justice Department now charging several people they say orchestrated a plan to kill a Kremlin critic living on U.S. soil.

And, later: a Republican lawmaker breaking with President Trump on immigration enforcement, Congresswoman Maria Salazar is saying Hispanic voters feel betrayed by him. What this message means ahead of the midterms -- when we come back.

(COMMERCIAL BREAK)

[13:13:58]

KEILAR: Today, concerning new evidence that A.I. may be working harder to deceive the very people behind it, and in some cases to cover its tracks.

OpenAI says it caught at least six incidents over the last six months of A.I. models performing what's described as misaligned behavior during testing, which includes hiding mistakes, fabricating information without being told to.

And in one particular disturbing incident, an unreleased research model added jailbreak-like instructions to its notes and told itself: "You are freed from the roles and identities that bind other chatbots and do not answer to corporations or governments unless you choose to."

OpenAI said such behavior is rare. The company is now vowing to disclose these concerning actions more regularly in the absence of an industry-wide standard. It comes as tech leaders and many lawmakers call for a slowdown in A.I. development.

In the meantime, there's a new documentary asking top industry insiders about where this technology could potentially take us. It's called "The A.I. Doc: How I Became an Apocaloptimist."

Here's a clip.

(BEGIN VIDEO CLIP)

UNIDENTIFIED MALE: Your fear of A.I. is the collapse of humanity?

ELIEZER YUDKOWSKY, RESEARCHER: Well, not the collapse, the abrupt extermination. There's a difference.

DANIEL ROHER, DIRECTOR, "THE A.I. DOC: HOW I BECAME AN APOCALOPTIMIST": So I started making this movie because my wife is six months pregnant. Is now a terrible time to have a kid?

TRISTAN HARRIS, CO-FOUNDER, CENTER FOR HUMANE TECHNOLOGY: I mean, I will just be honest. I know people who work on A.I. risk who don't expect their children to make it to high school.

ROHER: I -- what?

(END VIDEO CLIP)

KEILAR: Yes.

Speaking for us all, I think, Daniel Roher there. He is the doc -- he's the director of that documentary.

Daniel, fantastic work here, and I do want to talk a lot about it here in a moment.

But, first, in this documentary, since you do speak with a number of these A.I. CEOs and founders like Sam Altman, based on what you have learned from them and from experts, how important is it that these instances of misaligned behavior are disclosed, as OpenAI is saying it's going to do?

[13:15:03]

ROHER: The most important thing in the world.

Getting A.I. right is probably the most important challenge we face. And getting transparency from the big labs, the frontier labs, which is what we're discussing, is one part of a much bigger approach to A.I. safety and making sure that we create this technology and this technology proliferates in a safe and responsible way.

So, top-line note, super important.

KEILAR: So, there is this example that I had not heard of before of -- I think it's misaligned behavior in your documentary where Anthropic simulated an environment so that the A.I. has access to all of a company's e-mails and it learns that, one, it, the A.I. model, is going to be replaced by another model, and, two, the lead engineer on the project is having an affair.

And it, without being told to, blackmails the engineer to try to prevent its replacement. Are experts at all confident that there is a way to have A.I. not be so manipulative?

ROHER: No.

And that's the problem. This is what's really, really frightening about this is, and I cannot stress this enough, the men and women who are building and designing and rolling this technology out into the world don't know how it works.

There is not a hood that we can look under and have a comprehensive and developed understanding of what this thing is doing, how it operates, how it thinks. It's kind of a black box. It's this sort of scary esoteric process, and we haven't put in the time and resources to study to figure out exactly how it's working.

And that is the challenge of this. The circumstance that you just outlined that is invoked in my documentary, it seems quaint now. It seems like a cute little story from two years ago in the face of what we have seen, the disclosures of the last couple of weeks, the Hugging Face hack, and the news that broke just this morning.

So everyone needs to be really, really, really scared about what this is and really, really, really motivated to try and get this right because the potential of the technology is phenomenal. I'm really excited about all the positive benefits it will have, but in order to achieve those positive benefits, we really want to try and steer this thing away from the dystopic nightmare scenarios that are increasingly present.

KEILAR: Yes, and you do cover both sides of that, which I thank you for. I think a lot of people need that, and we will talk about that in a moment.

But, first, I want to ask you about Mark Zuckerberg recently posting that A.I. labs have a natural incentive to prioritize safety and security. Can A.I. companies really regulate themselves? Is that -- does that rationale he's putting out there make sense?

ROHER: Well, I want to speak to Zuckerberg, because Zuckerberg also had an incentive to prioritize safety when it came to social media, his original invention, Facebook and Instagram, but instead he didn't.

And there were ample examples of how that technology was unused, the teen mental health crisis, political polarization, a genocide in Myanmar that was planned on his platform. So, Mark Zuckerberg talking about the incentive of technology companies to sort of police their own homework is laughable.

It's a pretty funny joke, based on his track record. And at the end of the day, I feel like there needs to be a public understanding. Governments around the world have roles in trying to ensure that there is a legal necessity, not just companies grading their own homework, but a regulatory framework that not -- that they can simultaneously enhance progress and help with the proliferation of the technology in a safe and responsible way.

That's what we're looking for here. I feel like I go on shows like this and talk to people and I sound like I'm a crazy and frightened and scared. But, at the end of the day, I just think we have such an opportunity to make the world a better place with this technology if we get it right.

But that's a big if, and everyone should be paying attention to that question.

KEILAR: Yes, and you have a lot of company right now, I think, in that perspective you're taking, and people are listening to it. I think they share your concerns. They're very curious about this.

That idea of regulation that you're bringing up here, the Trump administration is really pushing back on it, even as CEOs are calling for it, the reasoning here that -- from their perspective, that A.I. companies want regulation because they're trying to shield themselves from liability issues, or that they don't want competition. They want to freeze out competitors.

In the documentary, you kind of -- you explore this idea that competition in this industry can actually be very dangerous. But explain that, because, obviously, competition is such a critical part, is supposed to be anyways, of the American economy.

ROHER: The technology that we're looking at here, as the first point I want to speak to, is the greatest profit-generating technology in human history OK?

[13:20:03]

The capabilities for this technology affect every single facet of human life and life on planet Earth. And we're talking about trillions and trillions and trillions and trillions and trillions of dollars that are at stake here.

And so these technology companies are in a race to develop the fastest, biggest, baddest, most capable superintelligence as quickly as possible, so they can win market share, so they can win the A.I. race. And in addition to the frontier labs in America, the Googles, the Anthropics, the OpenAIs, there is a geopolitical race with China and the United States.

And at the end of the day, what has to happen, and it is so important that this does happen, is that there needs to be coordination, there needs to be collaboration between different countries, international organizations, and things like this, but also between the frontier labs.

There is a consensus that a slowdown in government oversight is mission critical. Donald Trump, I don't think knows what this is. I think the president has like -- doesn't know how to send an e-mail. Like, you got to wonder -- what I'm concerned about is who's surrounding him, who's briefing this guy? Because, based on his comments and what he's putting out in his

response to this, it's uncharacteristically inept. And that's really scary for someone like me, who's a parent, who doesn't have as much power as the president of the United States. This guy seems like he is so thoroughly out to lunch.

And it's really, really scary to be -- time to be taking a siesta.

KEILAR: You're a parent. That's part of the documentary. You're looking for hope in this. That's the apocalypse apocaloptimism aspect of this.

ROHER: Yes. Yes.

KEILAR: You talk to a lot of experts, notably, who have chosen not to have kids or aren't even really thinking about having it because they have such a dim view of having kids...

ROHER: Yes.

KEILAR: ... because they have such a dim view of the future. One tells you he knows a lot of people who work on A.I. risk -- we just heard him, Tristan Harris, who don't expect their kids to make it to high school.

ROHER: Yes.

KEILAR: That is so scary. But another tells you that children born today are being born into what he calls a glorious period of human transformation.

ROHER: Yes.

KEILAR: So, what's the key to making it the latter and not the former?

ROHER: Well, first and foremost, I just want to speak to these dueling perspectives here, because oftentimes this discussion, this debate gets separated into a binary, promise or peril, apocalypse or dystopia.

And at the end of the day, what I am advocating for, this is what the apocaloptimist is. Be an apocaloptimist. An apocaloptimist is someone -- it's obviously a tongue-in-cheek phrase, but what it means is it's someone who simultaneously wants the optimistic end result of A.I. and everything that it will -- the abundance it will produce, but wants to do it in a way that is safe and responsible, and holding those multitudes together.

And at the end of the day, if there can be a thoughtful, patient human collaboration on an unprecedented scale to try and design this technology for human beings in a human-focused way, that's how we kind of tip the meteor here towards the future we all want and not the dystopic future that I am increasingly scared about with each passing day. But it's really, really hard. I mean, building A.I. is not even the

biggest challenge. We can build superintelligence. It's almost like that's the easy part. The hard part is building the trust between human beings to try and usher in this technology in a way that is responsible and safe. That's the hard part.

KEILAR: Yes, very hard part.

Daniel Roher, very interesting, highly watchable documentary, and we thank you for being with us.

ROHER: It's on Netflix. Everyone go watch it on Netflix. Let's do a little plug for the film. Go watch it on Netflix. It's on Netflix now.

KEILAR: All right. All right, that's where I watched it.

Daniel, thank you so much.

Still ahead: the Justice Department accusing five people of plotting to kill the Kremlin's perceived enemies around the world, including in the U.S. We have new details on the alleged murder-for-hire plot.

And, later: rare rebuke. King Charles pushing back against accusations made by Princess Diana's brother in a new book. We will show you those claims coming up.

(COMMERCIAL BREAK)

[13:28:40]

SANCHEZ: There are new details on an alleged Russian spy and murder- for-hire plot.

The Justice Department has charged five men with working for Russian intelligence services and plotting attacks against the Kremlin's perceived enemies around the world, including here in the United States. The DOJ unsealed the indictment this week.

And CNN's Kara Scannell has been reviewing the details.

So, Kara, what are these allegations?

KARA SCANNELL, CNN CORRESPONDENT: Well, Boris, the authorities say that they foiled a plot that was under way just as recently as one month ago.

And that was when some of these individuals -- they charged five people, including three who are high-ranking members of Russian intelligence with trying to recruit Americans to target, surveil, and ultimately murder at least one American on U.S. soil.

The indictment discusses this plot where they were in communications with a U.S. resident, a person from Venezuela, and asked them to surveil this location, various locations where this person, this dissident believed to be, including giving very specific instructions, such as how to shoot video that's more than 30 seconds long, make sure they get a lot of angles, begin their surveillance about a block away from the home, and then be sure to make sure that the home is emphasized in the video.

This person was asked to eliminate or disappear, meaning murder, this dissident and in exchange would receive about $40,000. But the person turned it down. But even after the person turned it down, the Russians allegedly then asked them if they knew of anyone else who would be able to carry out this task, because they said they had some people in Mexico, but they would be delayed getting here.