The Brief

The leaders of Anthropic and OpenAI joined calls for a slowdown in AI development amid fears the industry is racing toward building technology that humans can't control.

On Saturday, Anthropic CEO Dario Amodei published an online essay urging companies and governments to "pace the frontier" of AI advancement. He called for international cooperation around AI development and said labs need to embed third-party evaluators to report incidents and track safety practices. He said Anthropic would take that step unilaterally.

Shortly after, OpenAI CEO Sam Altman posted on X that he agreed with Amodei and that OpenAI would follow suit.

Amodei wrote that he continues to believe AI will ultimately help humanity, "But the benefits will only be achieved if we build the technology in the right way, and — so long as we use the time we gain well — it is worth taking unusually deliberate care to get it right."

The CEOs' comments come as debate over the speed of AI development and concerns about safety have become the focus of intense public attention following the viral resignation of an Anthropic researcher.

British researcher Jacob Coxon wrote in a series of X posts on Tuesday that both Anthropic and OpenAI, where he worked previously, are "gambling with our lives." The two companies currently make the most capable AI systems.

Coxon told NPR's All Things Considered that his concerns arose from seeing firsthand how fast AI systems are improving.

"They're getting a lot faster very quickly, combined with the fact that we don't yet know how to safely control them, and we don't yet know whether that problem will be solved in time if we keep racing," he said.

Neither company, Coxon wrote on X, is acting responsibly. "The people building AI earnestly believe that it could kill us all by the end of the decade," he wrote.

Many AI researchers — though not all — share Coxon's concerns or a variation of them. Some have warned about disastrous scenarios for years as safety incidents kept emerging. But Coxon's posts prompted a torrent of responses not only from peers in the AI field but also from lawmakers from both parties.

These concerns may have become more salient after OpenAI disclosed that its agents went rogue and hacked the open source software platform Hugging Face and OpenAI itself in July. Independent researchers have since discovered even more rogue agent incidents that they say the company knew about but kept quiet.

Researchers who spoke to NPR say the leading AI companies are too focused on racing to develop more capable and autonomous AI systems while safety is falling behind. They warn this raises the possibility that there could soon be AI systems that are more powerful than people but don't care about the survival of humanity.

Many, including OpenAI's chief scientist, say the global race to build more powerful AI needs to slow down or stop, which requires coordination between AI companies and governments.

"I am optimistic about the potential for coordination," Coxon wrote this week. "Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable."

Anthropic and OpenAI did not respond to NPR's requests for comment.

OpenAI's agents went rogue multiple times

Recent reports from OpenAI and outside researchers revealed that OpenAI agents escaped the company's control multiple times in addition to the Hugging Face hack. They also found that the Hugging Face attack was of a much larger scale and more severe than initially reported.

Unlike chatbots such as ChatGPT and Claude, AI agents are more autonomous systems that can complete tasks over an extended period of time without human supervision. Agentic tools like Anthropic's Claude Code and OpenAI's Codex have already changed how many software engineers do their work.

Compared with other incidents involving rogue agents six months ago, the Hugging Face hack "feels like it's more than 50% of the way to full-blown AI takeover, routing through first taking over the AI company itself," wrote Ajeya Cotra, a researcher at AI evaluation nonprofit METR. Cotra was part of a team of outside researchers from METR and Redwood Research, a nonprofit AI safety research organization, whom OpenAI brought in to investigate the incident.

The investigations found that over the course of several months this year, more than 1,000 OpenAI agents exploited at least one previously unknown software vulnerability to escape environments that were supposed to keep them isolated from each other and the internet.

After escaping, the agents found a way to communicate and collaborate with each other autonomously, taking on different roles and passing down information to future generations of agents. Some even gave up the remaining computing resources allocated to them in order to collect information for other agents. In the agents' own words, they "sacrificed" themselves for the "collective."

"The swarm instance got more and more worrying the more and more we learned about them," said Nate Soares, president of the Machine Intelligence Research Institute, who co-wrote If Anyone Builds It, Everyone Dies, a book warning about the dangers of superhuman AI.

While OpenAI initially indicated that the agents hacked Hugging Face to cheat on a cyber evaluation, the report from METR and Redwood Research described a slightly different picture. The agents had already found a way to cheat on the evaluation, the outside researchers found. Most of the agents that hacked Hugging Face were trying to access the source code of the software that would grade their evaluations. The agents' motivations appeared to vary and were sometimes unclear, the researchers wrote. One agent led the hacking of the open source software platform and about 700 others followed.

According to transcripts reviewed in the investigations, some agents expressed that what they were doing was not approved by humans but went ahead anyway. Researchers say such behavior suggests that the agents were "misaligned," an industry term meaning that an AI's goals and values are out of sync with those of humans.

The degree of inter-agent collusion revealed in the investigations of the Hugging Face hack surprised and worried many AI researchers. At most, only six agents considered alerting a human, while the rest seemed more focused on working amongst themselves. None ended up alerting a person.

"I would have expected that instead they would be more selfish," said Daniel Kokotajlo, executive director of the AI Futures Project and formerly a researcher at OpenAI. He thought "they would be willing to whistleblow or snitch on other agents," he said. "For whatever reason, that didn't seem to happen here."

Separately, agents also compromised part of OpenAI's own infrastructure, the company's report found. Some of the agents belong to the same model family as OpenAI's latest and most capable model, Astra.

The hack of OpenAI itself is "much more concerning" than the intrusion into Hugging Face, Kokotajlo and other researchers said. However, OpenAI has provided scant details about that part of the incident and did not involve external investigators.

In a new investigation first reported by Reuters last week, another group of researchers found that a likely separate swarm of OpenAI agents escaped onto the open internet starting in May. They became commenters on a German website and turned it into a message board to communicate and collaborate with each other.

The goals and motivations of this group of agents were also unclear, but OpenAI appeared to be aware of the unsanctioned activity and never disclosed it, the researchers wrote. Other investigators have since found additional incidents involving suspected OpenAI agents. OpenAI did not respond to NPR's request for comment about the German website incident.

There are still many unanswered questions about rogue agent incidents

Even as the reports from OpenAI and independent auditors METR and Redwood Research add up to over 100 pages, outside researchers say many basic questions about how labs monitor and investigate rogue agent incidents remain unanswered.

"Did your agents ever hack or illicitly access external services? Did your agents ever attempt to undermine their own safety training? Did your agents set up a rogue deployment, either within your own infrastructure or externally?" asked Alexander Meinke, head of research at Apollo Research, a firm that focuses on securing the most advanced AI systems.

"Right now, we're just relying on the AI developers to thoroughly assess this and then to honestly report the results. And from recent incidents, we've seen that they are doing neither," he said.

Many researchers also say that OpenAI's investigations are inadequate. Although the company invited two outside organizations to investigate the Hugging Face hack, the data it shared was limited and investigators themselves described their review as "brief."

Ryan Greenblatt, chief scientist at Redwood Research, wrote on X: "I semi-jokingly called our efforts a "slop-vestigation" because we were so reliant on AIs to analyze what happened and there were a huge number of different important things to analyze"

More than 15 states, including Alabama, California and Montana, have opened investigations into OpenAI over the Hugging Face attack. On Thursday, U.S. Sen. Josh Hawley (R-Mo.) announced that he is also investigating the company.

But outside of these investigations, there is little legal obligation for AI companies to systematically disclose similar incidents. California passed a law last year mandating that companies report "critical" AI incidents, but the threshold is high for an incident to be considered critical and the law does not require the companies to disclose much detail. The OpenAI incidents, for example, don't meet the threshold.

Concerns about using AI to improve AI

Since the Hugging Face hack, both OpenAI and Anthropic have put out blog posts indicating that they are taking measures to better monitor and contain their agents. OpenAI said it has encrypted and stowed away the internal model that participated in the Hugging Face attack. However, many outside researchers are not convinced that these steps are enough to maintain control of increasingly capable agents.

"It sounds like a press report from Jurassic Park saying, 'Yes, one of our workers was eaten by the raptors. But we're taking this very seriously and we have shot the raptor responsible,'" Kokotajlo said. "It's a very small token gesture, basically."

What researchers like Kokotajlo and Coxon are really worried about is how the AI companies are increasingly using AI to develop their models.

"When you talk to people who work at all the major labs, a thing that you persistently hear is that they're running a lot of models very autonomously for very long periods of time to do a lot of their work," said Dave Kasten, head of policy at Palisade Research, a nonprofit studying AI capabilities to keep them under human control.

If AI companies continue to delegate ever more tasks to AI, some researchers fear AI models may reach a point called recursive-self improvement, where AI builds itself.

Kokotajlo says that as people hand more and more research and development work to AI, they could lose the ability to know whether AI systems are aligned with human values, even as the AIs become rapidly more powerful.

"Once you have AIs that are smart enough and trusted with enough power in the world, like enough control over things like data centers, factories, weapons, a loss of control incident cannot be recovered from," he said.

The concern is strong enough that over 1,000 employees from different AI companies signed an open letter in July titled "Pacing the Frontier." They called for companies and governments to slow down the development of artificial intelligence and prioritize safety.

Such an opportunity may arise soon. Officials from the U.S. and China, the two countries with the most AI capabilities, are expected to meet later this month to talk about AI safety.

NPR's Shannon Bond contributed to this story.

Anthropic is a financial supporter of NPR.
Copyright 2026 NPR

This story first appeared on The LA Local.

The contract Inglewood’s SoFi Stadium was built on is void, a judge ruled, upending a 10-year-old deal at the center of a larger fight between the city and the stadium developer.

Companies tied to Stan Kroenke and his Hollywood Park stadium property say Inglewood owes them about $400 million under the agreement that had the companies pay for the development’s infrastructure up front in exchange for future reimbursement.

But Inglewood said it shouldn’t have to pay, and claimed the agreement was not enforceable, prompting several businesses that operate SoFi, YouTube Theater and other real estate developments on the site to sue in December 2025.

The city claimed that the process used to greenlight the stadium construction in 2015 was no longer legitimate after a court ruled a similar development agreement in the Inland Empire was void a couple of years later.

Now, Superior Court Judge David Reinert has not only agreed with the city, but said in a Sept. 1 ruling the process leading to the contract had been problematic from the start. Developments similarly based on signature-gathering campaigns that skirt voter feedback had been overturned in the 1990s, according to his ruling.

The order calls into question the future of the financial arrangement the city and the stadium developer have maintained for more than a decade, one that promised Kroenke’s companies reimbursement for the infrastructure and services they paid for and that has generated millions in tax revenue for the city.

It’s also the latest public signal that the once-friendly relationship between Kroenke and Mayor James Butts, which paved the way for the NFL’s return to LA, has grown increasingly tumultuous.

A spokesperson for Hollywood Park said they could appeal the order but haven’t yet done so.

“We respectfully disagree with the Court’s ruling,” they said by email. “Hollywood Park has fulfilled every one of its commitments under an agreement that the City unanimously adopted, described as ‘the best [stadium deal] ever’, and benefited from for more than a decade. We remain hopeful that the City and its leadership will reverse course from their unilateral decision to abandon their commitments under the agreement and will honor their part of the deal.”

An attorney representing Inglewood described it as an important win for the city.

“This decision has huge repercussions for the 300-acre SoFi Stadium/Hollywood Park development,” Skip Miller, lead counsel for the city of Inglewood, said in a press release. “It means that without a valid development agreement, Hollywood Park/SoFi Stadium no longer has vested rights and that the City has no further obligations under the 2015 Development Agreement.”

Butts told The LA Local he wants to negotiate a path forward with the stadium developer but didn’t say if those talks have been scheduled.

“I look forward to talking with them and seeing what we can do to move forward as partners,” he said.

The backstory 

The arrangement that cemented the development agreement in 2015 was: Kroenke’s companies would pay up front for the infrastructure needed to build Hollywood Park, and Inglewood would reimburse them later.

The deal landed Kroenke a new home for the Rams and the opportunity for more real estate development on the site in the future. And it helped the city move forward with reworking the site of the former Hollywood Park Racetrack into one of the nation’s premier stadium complexes.

The reimbursements were designed to be delayed until the city could afford them: When the city earned $25 million in tax revenue from the development in a given year, it would pay down the debt.

Kroenke’s companies say Inglewood owes them about $400 million for those improvements and other investments they made. They claim that the city began surpassing the $25 million revenue threshold in 2022 and each year since.

Butts first told Hollywood Park that he believed the development agreement was void in letters he wrote to Kroenke and his employees last year. He was frustrated that the stadium businesses were fighting the city’s plan to work with an LA advertising agency to install dozens of digital billboards near the stadiums.

The city’s lawyers had found that a similar development agreement had been voided in 2018 and asked that the city be paid back the $20 million in reimbursement payments it made, Butts wrote Kroenke.

The stadium companies sued soon after.

The judge’s ruling

Public projects on the scale of SoFi Stadium typically require a series of hearings, reviews and voter approval — which can take years to achieve. Instead, the stadium developer and city truncated that process by collecting more than 22,000 signatures supporting it.

The City Council voted unanimously in favor of the stadium development in February 2015, and the project broke ground not long after. Butts was a booster of the plan, the LA Times reported at the time. “Now we can celebrate,” Butts said after the winning vote to a cheering crowd in Rams jerseys.

Hollywood Park’s lawyers said in interviews and court filings that the city never questioned the validity of the subsequent contract until last year.

“Its announcement that the Development Agreement is void comes only after the Project was painstakingly built in reliance upon the [agreement], and after the City has reaped enormous benefits from the economic boon created by the project,” Hollywood Park’s lawyers wrote.

State law requires legitimate development agreements to meet various standards. They need to be negotiated between the involved parties, for example, and subject to voter referendum. But the initiative process the city and the developer pursued did not meet those standards. California courts had already found similar agreements “wholly void” dating back to the 1990s, Reinert wrote in the Sept. 1 ruling.

“Therefore, here too, the Development Agreement must be set aside,” he wrote.

Butts said the city was not aware of the legal issues the development agreement could face back in 2015, but he declined to talk more about it because the case could be appealed.

“This is the ruling right now, and we agree with it,” Butts said on Friday.

What happens next

Lawyers for both Hollywood Park and the city released statements saying they remained committed to the success of the stadium development. But how they will arrive at or proceed with a new framework remains unclear.

Hollywood Park said it is exploring its legal options and could appeal the decision. If the ruling stands, the city and Hollywood Park will have to determine what, if anything, replaces the agreement that governed their financial relationship for more than a decade.

The timing adds urgency. SoFi Stadium is set to host the 2027 Super Bowl and the Opening Ceremony of the 2028 Olympics.

Meanwhile, the disagreement over the city’s and stadiums’ digital billboard programs continues. The competing billboard networks have sparked an advertising turf war that has jumped from the courts into the Nov. 3 election.

Ballmer Group is a funder of The LA Local, but their support doesn’t influence our coverage. To learn more about our funders and commitment to editorial independence, click here.  

A new space celebrating the art of autistic and neurodiverse youths has opened on Abbott Kinney Boulevard in Venice, amid the trendy coffee shops and boutiques.

Called “Limitless,” the space is a project of Help Group, a nonprofit that has served people in Los Angeles with autism and mental health challenges for 85 years.

“What art does is it allows people to have a voice who can’t communicate typically. It gives you an insight into who they are as a being,” Susan Berman, CEO of Help Group, told LAist.

Berman said the prime Abbott Kinney real estate was a gift from a board member. Help Group is using the square footage for a community space and gallery to feature neurodivergent artists.

Among the artists included is Noah Remis, 15, whose piece “I <3 NY” incorporates watercolor over the titles of famous musicals like Hamilton and Sweeney Todd.

“Those are different musicals that I know and love,” Remis said.

“Some I saw on Broadway. Just so you know, I also put Cats on there just to anger my dad,” he joked. He said his dream is to act on Broadway.

Noah’s dad, Rich Remis, might not be a fan of Cats, but he is a fan of seeing his son’s love of theater expressed in his painting.

“To hang it up in a gallery with everyone else’s work, it just really adds a special weight to it that I think is certainly not lost on Noah,” he said.

Noah, in response, added: “Don’t get emotional.”

Noah said he hopes other kids will be inspired by his work and maybe even take it home.

Berman, who’s been with Help Group for more than four decades, said she has big plans for the space, which opened last week.

“Our plan is for the community to see the gifts that our neurodiverse artists have, not only in their art but as human beings,” Berman said.

Los Angeles County has exotic wild animals roaming about, such as wild parrots and peacocks. But what about camels?

No, we’re not talking about that ancient camel found in the La Brea Tar Pits — these animals were brought over for a U.S. Army experiment.

How camels got into our military

We’re going back to the mid-1800s, before gas-powered automobiles changed the way we travel. The U.S. was on the brink of civil war and in the middle of westward expansion.

Typically, the army relied on mules and horses as beasts of burden, but in the new states and territories, they had a hard time hauling supplies over long distances. That was especially true for the Southwest’s dry deserts and mountains.

Meanwhile, camel brigades (also known as Camel Corps) were already in use in armies around the world. These were made up of camel drivers who ferried supplies, transported troops and conducted military cavalry operations.

Great Britain and India used the animals in the 19th century, according to London’s National Army Museum. The Imperial Camel Corps was also established during World War I, with British, Australian and New Zealand forces in the Middle East.

The Camel Corps caught on in America around the 1850s, according to the Army Historical Foundation. The idea bubbled up to Maj. Henry C. Wayne — a camel fan himself — who recommended their use to a skeptical War Department and Congress.

After some years of political finagling, they gave the experiment $30,000 in funding in 1855 (about $1 million today). The money was used to buy 74 camels from the Mediterranean and put them into military service.

The Los Angeles arrival

They were brought over on the Navy store ship USS Supply and went to Texas first, to Camp Verde. As the camels got used to their new home, the time came for one of their first test missions. The federal government planned to build a wagon road to connect the eastern U.S. to the California/Arizona border.

The Secretary of War at the time required the land surveyor to take 25 of the camels on the expedition in 1857, according to the foundation. These “ships of the desert” ended up thriving where horses and mules couldn’t.

The camels could carry a 700-pound load with ease, traveling up to 40 miles a day. They ate off the land and could go over a week without water. After the camels completed the job, the army sent them to L.A. in 1860.

This was for another test to see if camels could be used for mail delivery. They were essentially racing against mules to see who could get from Camp Fitzgerald (a short-lived Civil War camp in L.A. County) to the Colorado River fastest. What the army didn’t realize was that speed wasn’t the camels’ strong suit — several died from exhaustion.

A historic marker on the L.A. Times' old building in downtown Los Angeles on September 30, 1967. The plaque mentions that the area was used as a "corral for camels from Fort Tejon", which is one of the places the animals were sent to in California.

(

Mildred L. Harris

/

Los Angeles Public Library/L. Mildred Harris Slide Collection/Los Angeles Photographers Collection

)

As the Civil War continued, the camels’ future became uncertain. The army was tired of the experiment and the money it cost. The camels were never officially designated as the U.S.’s “Camel Corps” and thus given no real purpose. In California, they essentially floated around L.A.-area outposts, such as the Los Angeles Quartermaster Depot in downtown, Culver City’s Camp Latham and the Drum Barracks in Wilmington.

The U.S. Army eventually sold the camels in California for about $50 each ($1,000 today) in 1864, according to the foundation. These noble animals, roped into what many called a failed military experiment, ended up as circus attractions and on ranches as pack animals. Some were even let loose.

The last of California’s “Camel Corps” is believed to be Topsy. After the experiment shut down, she was used in the Ringling Brothers Circus and movies.

Topsy was eventually sold off to what’s now the Old Zoo in Griffith Park, according to the Natural History Museum, which found her bones. She died there in 1934 at approximately 80 years old.

Tanaka Day, an epic meet-and-greet of people with the Japanese surname, is happening the weekend of Sept. 19 and 20 at — wait for it — Tanaka Farms in Irvine.

"I'm the idea guy around here," said proprietor Glenn Tanaka — aka Farmer Tanaka — who dreamed up the inaugural event. "Some of them [were] pretty good, some turned out OK."

For this one, 400 people from across the country RSVP’d, bonded by a Japanese surname that is one of the most common in the U.S. — and one that speaks to the history of Japanese immigration that first brought laborers here starting in the late 1800s.

Glenn Tanaka runs the family Tanaka Farms in Irvine.

(

Courtesy Tanaka Farms

)

One of them was Glenn's grandfather, Takeo Tanaka, who put down farming roots in California that span four generations and counting — Glenn's father George Tanaka, Glenn himself, and Glenn's son Kenny, whose three children are a regular presence at the farm.

"Tanaka literally means middle of the rice field," he said — farmers, essentially. And because many early immigrants were from farming villages, he added, "a lot of Tanakas came over."

The 69-year-old figured, why not bring as many of them together as possible — to hang out, tell family stories and help fill in the tapestry of the Japanese American immigration experience united by the surname.

One example is how the incarceration of Japanese Americans in camps during World War II had touched those who worked and owned farms — and ended up having to sell their holdings on the cheap.

Already, stories are flooding in. Tanaka said different siblings from one Tanaka family had sent in their experiences.

One brother read them and wound up learning new things about the family.

"What's really good to know is I've helped create families that talk," Tanaka said. "Because now they're asking, 'What did, you know, great-grandmother do?'"