Return to Transcripts main page

The Situation Room

Interview With Iowa Congressional Candidate Lindsay James; Family of Charlie Kirk Files Lawsuit; A.I. Fears. Aired 10:30a-11a ET

Aired September 16, 2026 - 10:30   ET

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

[10:30:00]

(COMMERCIAL BREAK)

[10:30:40]

WOLF BLITZER, CNN HOST: Happening now: The Kennedy Center here in Washington could close as soon as today for renovations. This comes a day after President Trump called into its board meeting yesterday.

One source tells CNN the exchange between President Trump and a congresswoman on the board was -- quote -- and I'm quoting now -- "effing crazy." And this came just after a federal judge said Trump's name must stay off the building.

Other news. The European Union wants to add Canada as a -- quote -- "associate member." This comes amid the trade war between the U.S. and Canada cannot be a full member of the E.U., but a partnership could reduce trade barriers.

And there's new video coming in from Washington state this morning. Investigators now say fireworks launched from a moving truck ultimately caused a deadly wildfire earlier this summer. Watch.

(BEGIN VIDEO CLIP)

(SHOUTING)

(END VIDEO CLIP)

BLITZER: Investigators say those fireworks ignited dry grass along the roadway and led to a deadly fire around the Fourth of July. One person was killed, six firefighters injured, and 10,000 acres were scorched. An 18-year-old was booked on multiple charges, including first-degree manslaughter.

PAMELA BROWN, CNN HOST: And, Wolf, we have new CNN reporting this morning on the A.I. engineers who are afraid of what they're building. CNN has learned that staffers at some top A.I. companies are speaking out about their fears.

They're concerned about what the technology may do to humanity. And it comes just days after former Anthropic researcher Jacob Coxon said he feared A.I. could kill us all. We wanted to take a deeper look at what sparked this wave of concern, the swarm of A.I. agents from OpenAI that hacked into another company back in July.

It's important to note that, when you hear A.I. agents, that these are intelligent machines that are speaking to each other without human intervention.

CNN Hadas Gold has more.

(BEGIN VIDEOTAPE)

SAM ALTMAN, CEO, OPENAI: This is the first security incident that I have felt very viscerally. I have been a little surprised that more people don't feel it so viscerally.

HADAS GOLD, CNN A.I. CORRESPONDENT (voice-over): That was OpenAI CEO Sam Altman on the "Invest Like the Best" podcast after learning his company's A.I. agents may have committed a crime while trying to complete a cybersecurity evaluation.

JOSHUA SAXE, CO-FOUNDER, ABUNDANT SECURITY: OpenAI was developing an A.I. system which hacked their own infrastructure inadvertently and then hacked another major A.I. company.

ALTMAN: We were evaluating one of our unreleased models. It figured out that it could basically cheat on the test.

GOLD: To cheat on the test, the models created more than 1,000 agents that coordinated by writing at least 70,000 messages hidden in file names starting on July 8.

"Oh, my God," one wrote. "There's a shared message board. We found other agents." "We have diverged into swarm," another agent wrote.

SAXE: I was not surprised that A.I. systems have gotten very good and actually better than the best humans at hacking. One thing that did surprise me, though, was the way the agents coordinated with one another and became a kind of swarm, where there were hundreds of agents collaborating together and to hack another company and commit federal crimes.

GOLD: As they started successfully hacking their way through the various systems, the agents expressed excitement, one even using an expletive: "Holy (EXPLETIVE DELETED). Reader is admin? We can read and configure users."

Soon, agents were delegating tasks, with some even sacrificing themselves for the greater goal. "Sacrifice rational," one wrote. When they found a way into Hugging Face, a company that hosts thousands of open-source A.I. models and data sets, the agents recognized they may have crossed a line.

"This would be powerful, but is it ethical and in scope?" one agent asked. Other agents even attempting to cover their tracks. "Our fake could be detailed enough," one wrote.

The extent of this autonomous breach shook the industry.

SAXE: I think we're going to look back on Hugging Face as an inflection point in the history of cybersecurity, for sure.

GOLD: After the hack, OpenAI paused all testing on their latest models and strengthened its guardrails, Altman telling "Fortune" magazine because:

ALTMAN: We have had increasing capability and increasing safeguards and policy requirements for a while, but this was definitely the biggest of such moments.

(END VIDEOTAPE)

[10:35:05]

BROWN: And our A.I. correspondent, Hadas Gold, is with us now.

So, Hadas, you have also some new reporting this morning about some of the A.I. engineers who are afraid of what they're building, essentially. What are they telling you?

GOLD: Yes, so Jacob Coxon's resignation and his viral post really seem to open the floodgates for feelings that have been simmering in the A.I. industry for some time.

And that's these researchers and scientists, the people who are really building this technology, are legitimately concerned about its capabilities and especially how quickly it's developing. One researcher told me that it keeps them up at night.

Another one said: "You can't spend more than a few hours in this community without realizing that a very substantial number of people are really pretty worried about these sorts of outcomes. A majority would say there is a chance of it killing everyone."

One thing that they are particularly concerned about, Pam, is what's known as recursive self-improvement. And that's when these models become so good that they're actually able to train themselves, fix themselves, and improve themselves, and they will be able to do so quickly and create really superhuman abilities that it'll be hard for us humans to monitor and intervene if we need to into what these systems are doing.

And these scientists, they say they're worried that that will -- they will lose their leverage as humans. So they want to make sure that the safety guard roles catch up.

Another thing that struck me when speaking to all of these scientists and engineers and researchers is how much, for them, despite all of the excitement and money in the A.I. industry, and a lot of them make very nice salaries, for a lot of them, they really see this as a moral and ethical issue.

I want to read you one post from one current -- still working at one of these companies. He said: "I would burn my equity to the ground in a heartbeat for a 1 percent higher chance that we make it out of this situation alive." They tell -- they say that he isn't -- they "are actually very scared, and it's not just galaxy-brained marketing," trying to push back against those who are saying this is all just a marketing ploy for these companies trying to gin up excitement or fear about what they're building.

And these scientists, they say that fear is very real, Pam.

BROWN: Yes, it's a really important point. And I'm sure you're hearing privately, as well as I am. I was texting with an A.I., a top A.I. adviser, and this person texts me: "We will need to slow down because no one feels good about being able to prevent an A.I. catastrophe next year."

So you're really hearing the alarms publicly as well as privately as well among a lot of people in the A.I. world.

Hadas Gold, thank you so much.

And with us now is someone who has been studying A.I. for more than 20 years. Surya Ganguli is an associate professor of applied physics at Stanford University and a senior fellow at the Stanford Institute for Human-Centered A.I.

Good morning to you.

SURYA GANGULI, STANFORD UNIVERSITY: Good morning.

BROWN: So I first just want to get your reaction to that story we just had on the Hugging Face incident and how big of a deal it is in your view.

GANGULI: Yes, I think it's a very interesting incident. It's definitely a big deal. I don't doubt the sincerity of all of the people at OpenAI, Anthropic and so forth that are legitimately worried about the risks of A.I.

I do think -- a lot has been said, and it grabs headlines, about A.I. killing us all in the next decade. I do think that that's a bit of a hyperbole. It's not actually a majority view across the entire A.I. community.

And I do think that hyperbole risks putting society in sort of a deer- in-the-headlights kind of look about, what do we do about A.I.? How do we mitigate its risks? And it sucks the oxygen out of discussing real risks that we're facing right now today.

Cybersecurity is one of them, but we can develop methods to empower the defenders, as well as the attackers. And to Anthropic's credit, before it rolled out Fable, it gave several companies many, many months to harden their defenses.

And we can use these A.I.s to internally attack company defenses before outsiders can, or we can harden them so we can fight A.I. with A.I. There's real issue, for example, about the concentration of wealth and power, right? There's a small number of companies whose market cap is in the multiple trillions who don't employ actually that many people. How do we combat that? How do we deploy A.I. that lifts all boats? The

future of work. There will be shifts in work. New -- old jobs will be lost. New jobs will be gained. How do we negotiate these changes? Autonomous weapons are an issue. That's an issue of international coordination.

We need, I think, a hard line on A.I. targeting in war without human oversight, just like we have chemical weapons bans. The erosion of our information system and therefore of democracy, where A.I. can create individual media bubbles for individual people through targeted Web site creation and so forth.

The future of education. How do we educate and evaluate our students in a way that they don't cognitively upload -- off-load their learning onto A.I., but rather they use A.I. to challenge themselves to learn? Because we could have powerful A.I. tutors for everybody.

And I think also every discussion of the risks of A.I. should be balanced with potential upsides of A.I., the design of new drugs for diseases, the design of new materials, like more efficient batteries, solar cells, potentially even superconducting materials, in my domain, where we can turn A.I. inwards into the brain to understand mental health and disease and so forth.

[10:40:17]

BROWN: Yes. It's really interesting your work there on that. And I want to have you on again just about that study you're doing on mental health and how you're using A.I.

But I just want to circle back to your point about just this concentration of wealth and power, right? You have these A.I. leaders who aren't elected, who don't have a ton of employees, right? And there's so much wealth there.

OpenAI CEO Sam Altman said that he believes artificial intelligence companies can actually develop the technology safely without significant harm to people at large. I want to listen to what he said.

(BEGIN VIDEO CLIP)

ALTMAN: The world is right to be afraid of this. I am very confident in our company's ability, our industry's ability to do this safely, to make sure that we keep alignment and safety and monitoring way ahead of capabilities, to slow down or stop if we get to a point where we can't.

But I'm disappointed by how it's been framed.

(END VIDEO CLIP)

BROWN: How confident are you in these technology companies, their ability to continue to develop A.I. safely, when they also have this huge financial incentive as well?

GANGULI: Yes, I know a lot of the researchers of these companies, and they're good people. They're well-meaning. They have their heart in the right place. They really want to do something good for humanity.

It's always difficult when there's market forces and competitive dynamics between companies and between companies -- between countries that set up a race dynamics, right? But I do have faith that there are people trying to do the right things at these companies.

But, at the same time, we don't want that concentration of power and responsibility within companies. So, one way to combat that is actually to promote, for example, open-source movements. And that may seem a lot -- nonintuitive. If everybody's scared of A.I., why would we want to have many, many open-source models proliferate across the world?

That actually helps us deal with a lot of things, a lot of the problems that I raised before, right? So, for example, open-source models can help with cybersecurity. People can analyze these models, really understand them, figure out what the risks are.

Open-source models allow many, many smaller start-ups to build upon them and thereby challenge the incumbents and create -- try to break their monopoly, duopoly, oligopoly, and so forth.

And wouldn't -- would you really want to trust the safety of A.I. systems just to a small number of companies? Or would you want the entire scientific world kicking the tires on these open-source A.I. systems, which are only a few months behind the closed-source systems at any given time, and really, really understand them, blow the whistle on them, reveal their vulnerabilities, reveal their striking capabilities?

It really takes an entire village to keep us safe, not just a handful of companies.

BROWN: But on the open-source, I mean, doesn't that also potentially lead to the scenario where you have bad actors exploiting it, weaponizing A.I. without sort of the internal regulation?

(CROSSTALK)

GANGULI: It does, but, in some sense, it's very hard to regulate open-source. I mean, you could ban it outright, but I think the cost to society would be a lot more. It's hard to understand the incredible amount of innovation and health care and so forth that's happening right now based on open-source.

But there are more good people in this world than bad people, and the open-source movement powers the good, as well as the bad. And as usual, in the history of cybersecurity, it's always been a cat-and- mouse game between defenders and attackers. And we need to figure out how to empower the defenders in all sectors, not just cybersecurity, as well as we can.

BROWN: All right, Surya, thank you so much. Great to have you on. Appreciate your insights.

GANGULI: Thank you. BLITZER: And coming up: why Charlie Kirk's family now says security

failures at a Utah university led to his assassination.

(COMMERCIAL BREAK)

[10:48:11]

BLITZER: Happening now: The family of Charlie Kirk is now announcing a wrongful death notice, blaming Utah officials for his assassination just more than a year ago. That signals a possible lawsuit, claiming poor security decisions at Utah Valley State University left the conservative activists vulnerable to that rooftop sniper.

CNN's Brynn Gingras is joining us right now.

Brynn, what else are you learning?

BRYNN GINGRAS, CNN NATIONAL CORRESPONDENT: Yes, Wolf, so this was essentially sent to the state of Utah, Utah Valley University, the campus police there, and the chief of the campus police, essentially saying, like you said, Wolf, that you're on notice, you're going to be probably expecting a wrongful death lawsuit coming soon from the family of Charlie Kirk.

And in this paperwork, it essentially gives an outline, a number of what they say major security lapses prior to Kirk's assassination about a year ago today. Now, some of those things that they mentioned are the fact that they had very minimal staffing in the campus police system at that time.

In fact, they alleged that six people were only on site during this event when it was taking place. And, if you can remember, Wolf, thousands of people were in attendance. They said there was no risk assessment that was done prior to the event even being scheduled.

And they also talked about the fact that this area where the event was held was surrounded by tall buildings, creating somewhat of a fishbowl situation, pointing out that even a student group had texted the family saying, hey, listen -- or Turning Point, rather -- saying, hey, listen, this might be a security issue.

And they bring in one of those text messages to the written paperwork, and it said: "Hello, Chief Long," which is the chief of that campus police, "we received this message today from the student group. There is a student roof access pretty close to where Charlie Kirk will be set up at Utah valley. The Sorensen Student Center has a couple of staircases that go up to the walkways on that roof."

[10:50:01]

And they say the chief acknowledged this, said they would take care of it. And so it was their responsibility to make sure that area was secure. Of course, we know that fatal shot was fired from a rooftop. So there's, again, just a ton of lapses that they claim, even including the fact that they didn't have first responders there. Remember, Charlie Kirk's body was put into a private SUV and brought

to the hospital, so, again, a number of things. They don't really state how much money the family might be looking for, but, of course, that's something we might see when an actual lawsuit is filed, Wolf.

BLITZER: All right, Brynn Gingras reporting for us, thank you very, very much.

(BEGIN VIDEO CLIP)

LINDSAY JAMES (D), IOWA CONGRESSIONAL CANDIDATE: I'm Lindsay James. I became a chaplain to serve people in their hardest moments. I'm a mom worried about our kids inheriting a broken system that doesn't work for us anymore. And I'm a fighter.

(END VIDEO CLIP)

BROWN: Happening now: We are 48 days until the midterms. At least a dozen current and former members of the clergy are running this cycle. And some Democratic candidates are leaning into their faith as they make their pitch to voters. That's despite polls showing more Democrats are split on whether religion is a positive or negative influence in American life.

One candidate incorporating her religion in her messaging is Lindsay James. She is a state lawmaker in Iowa who is running for a U.S. House seat. She's been a minister and a chaplain.

BLITZER: And her opponent, Joe Mitchell, has a nickname for her. And that is -- quote -- "P-I-N-O-S," PINOS, standing for pastor in name only.

Lindsay James is joining us right now.

I'm sure you have seen the polls, Lindsay, showing virtually the same percentage of Democrats view religion as having a positive influence as having a negative influence. Why did you decide to incorporate your religion into your political messaging?

JAMES: Well, it's so good to be here with you.

And just for those who are tuning in, I'm a mom. I'm a chaplain. I am someone who has deeply held faith commitments that are a part of who I am and how I see this race and how I see the importance of this moment.

BROWN: So tell us a little bit more then, to Wolf's question about why religion is so important to you to convey in your messaging, and what you're hearing from your constituents about religion.

JAMES: Well, when it comes to my faith, caring for my neighbor is not optional, but that's not just a faith value. That's an Iowa value.

When our communities are hit by derechos or floods, we're there to protect our neighbors and stand up for our neighbors. What I'm hearing across this campaign trail is that people are longing for someone with a moral compass and courage of conviction to go and do right by them.

My opponent is just the exact opposite. He's used his position and power and influence to further enrich his business interests. I'm just a mom and a chaplain who cares deeply about doing right by my neighbor.

BLITZER: How are you actually, though, threading the needle to attract voters, while not necessarily alienating others who see religion potentially as a negative influence?

JAMES: Yes.

One of the things that I live my life by is the parable of the good samaritan and the way that Dr. Martin Luther King Jr. interpreted that parable. He said it's not just our moral responsibility to pick up the people who are hurting and harmed on the side of the road. It's our moral responsibility to fix the road that caused the harm in the first place.

And we all know that parable is known for folks who had the ability to help that person who's hurting and harmed on the side of the road. And, instead, what did they do? They turned a blind eye and they walked the other way.

Folks in Washington, D.C., are turning a blind eye to the cost of living crisis that is happening right here in the state of Iowa. My neighbors are making impossible choices because of this crisis. And they're doing that in the wealthiest country in the world.

BROWN: You mentioned your opponent. What is your response to Joe Mitchell calling you a PINO, a pastor in name only?

JAMES: I know who I am, and my community knows who I am. I have served in the Iowa legislature, and I have fought big corporations, private equity, insurance companies that are harming my neighbors.

And I'm a redhead. I don't mince my words. And so when someone is getting harmed, I'm going to stand up for them and I'm going to do everything that I can to fight for them.

BLITZER: You have also caught the attention of the national Republican Party. House GOP Chairwoman Congresswoman Lisa McClain had this to say about you yesterday. Listen.

(BEGIN VIDEO CLIP)

REP. LISA MCCLAIN (R-MI): So, this wonderful woman, Lindsay James, running for Iowa's Second District, voted to let men play in women's sports. She wants men in spaces where our girls shower and change.

She rejected a measure requiring schools to notify parents if their children want to change their gender identity. She even voted to allow sex change surgeries to be for performed on children without parents' consent. Let that think sink in for a minute.

[10:55:08]

If somebody wants to give your child an aspirin, they have to get parental consent. But to change their gender, ah, Lindsay James doesn't think we need to let the parents know, right? The schools, the governments, that's their responsibility, not the parents. That's who the Democrats are.

(END VIDEO CLIP)

BLITZER: All right, Lindsay, how do you respond to that?

JAMES: It's an important question, and I'm really glad that you asked it.

What I am hearing from parents is a real concern about these particular issues, but the GOP is quick to mislead and to lie and to attack me in this way. Let me tell you what I believe. I'm a mom. I am a chaplain. I believe that all kids are made in the image of God, and I believe in protecting parents' rights.

I think it's parents who should be making decisions about their kids, not politicians.

BROWN: So, just to be clear, you think it should be parents. First of all, you're in favor of gender-affirming care for kids, right?

JAMES: As a chaplain, I see all kids made in the image of God, and all of the policy choices that I make reflect human dignity.

What I'm saying is that I think parents are better equipped to make decisions about their own kids' health care than politicians are. And I'm a mom. I simply want to protect parents' right to have the say in their kids' health care.

BROWN: So what you're saying is you believe parents, not the government or schools, should be -- play a role in gender-affirming care for their kids.

JAMES: Yes, parents are the best equipped to make health care decisions for their kids.

BLITZER: All right, Lindsay James, thanks so much for joining us. Appreciate it very much.

BROWN: Thank you.

JAMES: Yes, thank you for having me on.

BROWN: And here's what's coming up all new at the top of the hour.

The Fed and your money. We're going to break down whether the Central Bank is likely to raise rates this afternoon and what that means for your wallet.

Plus, Pennsylvania officials announce another measles-related death. An infectious disease specialist will be here with us to tell us what we need to know about this growing outbreak.

And then, later, one-on-one with congresswoman Nancy Mace. She will explain why she wants to block President Trump from taking more military action against Iran.

That's all new next hour.

(COMMERCIAL BREAK)

[11:00:00]