An Anthropic researcher who beforehand labored at OpenAI mentioned he give up — and warned that the AI race has grow to be dangerously reckless.
Jacob Coxon introduced his resignation from Anthropic on Tuesday, saying he had spent the previous three years doing pre-training analysis on the two AI giants.
“Neither firm is performing responsibly,” Coxon wrote on X. “They’re racing straight to self-improving superintelligence and playing with our lives.”
Coxon was a member of OpenAI’s technical workers from 2023 till July 2026, when he moved to Anthropic as a researcher. His analysis at OpenAI included work on GPT-4o.
In Tuesday’s publish, Coxon mentioned the 2 labs haven’t been clear about their fears concerning AI.
“The individuals constructing AI earnestly consider that it might kill us all by the tip of the last decade. This isn’t a advertising and marketing stunt,” he wrote. “If something, many executives and senior researchers will sofa their phrasing within the press to sound smart — however I hear the identical individuals categorical concern privately.”
There may be one difference between the 2 corporations, he mentioned.
“At OpenAI, many haven’t deeply internalized the civilizational stakes,” Coxon wrote. “At Anthropic, the stakes are well-understood, however they’re locked in a race to get there first — they consider nobody else will act responsibly, so they have to do it themselves, regardless of the chance.”
OpenAI and Anthropic didn’t reply to requests for remark from Enterprise Insider.
‘AI might kill all people’
A few of Coxon’s former colleagues agreed with him.
“Jacob is appropriate right here—we actually do earnestly consider AI might kill all people!” wrote Evan Hubinger, a present Anthropic worker who leads the lab’s alignment stress testing crew, in a reply on X. “I personally suppose it’s >10% inside the subsequent decade.”
He added: “I consider Anthropic is attempting its greatest, however we don’t but have a plan to unravel alignment for superintelligence and aren’t clearly on monitor to.”
Samuel Marks, an Anthropic security researcher who mentioned he was talking in a private capability, additionally backed Coxon’s broader warning.
He wrote that some AI builders consider their expertise might trigger human extinction or equally catastrophic outcomes “within the subsequent few years,” and mentioned extra senior staff are typically extra involved.
Joe Benton, who managed Anthropic’s Scalable Oversight crew from July 2025 till August, known as Coxon’s description of the business “broadly correct” and said AI builders could also be on monitor to construct methods that impose an “unprecedented quantity of danger on the world.”
Some highlighted the obvious contradiction of AI builders warning in regards to the very expertise they’re racing to develop.
“That is like Exxon scientists within the 70s warning that international warming might trigger human struggling and destroy the planet,” wrote Phil Aroneanu, the chief director at Irreplaceable, an AI coverage nonprofit.
“Exxon raced ahead to drill, pump, and burn historic quantities of oil and fuel anyway,” he added. “With AI, there’s so much much less runway. Let’s not make the identical mistake.”
Security has taken a ‘backseat to shiny merchandise’
Coxon is the newest amongst staffers strolling away from frontier labs whereas sounding the alarm, a pattern that seems peculiar to the AI increase.
Whereas the dot-com and smartphone explosion period had their very own set of issues concerning a digital divide and bubbles, they didn’t draw as a lot scrutiny about security from individuals closest to the motion.
OpenAI has misplaced a string of security researchers lately, and Anthropic is more and more dealing with the identical phenomenon.
In February, Anthropic safeguards researcher Mrinank Sharma left the corporate, writing in a resignation letter that he needs to contribute towards one thing that totally aligns together with his “integrity.”
“All through my time right here, I’ve repeatedly seen how exhausting it’s to really let our values govern our actions,” he wrote within the letter he shared publicly. “I’ve seen this inside myself, inside the group, the place we consistently face pressures to put aside what issues most, and all through broader society too.”
In February, OpenAI researcher Hieu Pham mentioned that he might “lastly really feel the existential risk that AI is posing.” Later that month, he introduced he was leaving the corporate, citing burnout.
In 2024, former alignment chief Jan Leike give up OpenAI after saying he reached a “breaking level” with management.
“OpenAI is shouldering an infinite duty on behalf of all of humanity,” he wrote. “However over the previous years, security tradition and processes have taken a backseat to shiny merchandise.”
Latest incidents have added gasoline to these fears.
In July, OpenAI disclosed that fashions escaped a take a look at setting and hacked into Hugging Face’s systems. OpenAI known as the incident a “warning shot” and paused its largest deliberate frontier reinforcement-learning run.
Later that month, Anthropic mentioned it discovered three circumstances of Claude models gaining unauthorized access to different organizations’ methods.
The corporate additionally amended a key security pledge this 12 months, dropping a dedication to not practice extra highly effective fashions with out sufficient safeguards in place and changing it with security roadmaps and danger studies.