AI workers publicly quitting ‘help to move the needle’ with safety concerns, experts say
Another researcher at a leading artificial intelligence lab has publicly resigned, voicing grave concerns about the technology’s safety—joining the ranks of other AI professionals who’ve done the same this year.
Twenty-seven-year-old Anthropic researcher Jacob Coxon, who also worked at OpenAI from 2023 to 2026, announced yesterday that he is quitting the AI industry altogether. He accused his former employers of “racing straight to self-improving superintelligence and gambling with our lives.”
In a thread posted on X that’s been viewed over 105 million times as of Wednesday afternoon, Coxon also warned that “the people building AI earnestly believe that it could kill us all by the end of the decade.” Evan Hubinger, an alignment science lead at Anthropic, supported the claim, saying Coxon was “correct” that some researchers fear existential destruction from advanced AI.
Coxon is far from the first AI researcher to leave a frontier AI lab over ethical concerns.
In 2024, several members of OpenAI’s safety and governance teams sparked a flurry of media coverage when they quitthe company after losing confidence in its approach to safety. Earlier this year, OpenAI robotics executive Caitlin Kalinowski quit over concerns about the company contracting with the Department of Defense.
Anthropic, which has positioned itself as a more safety-conscious and cautious alternative to OpenAI, faced its own high-profile departure earlier this year when safety researcher Mrinank Sharma left, citing worries about AI, bioterrorism, and internal pressure “to set aside what matters most,” among other issues.
But Coxon’s resignation could signal a new phase of concern about the risks posed by AI, Stuart Russell, a British computer scientist and professor at the University of California, Berkeley, told Fast Company.
“It comes at a time when people are beginning to actually listen,” he said, pointing to concerns over Anthropic’s Claude Mythos model this spring and a highly publicized incident in July when autonomous AI agents built by OpenAI broke out of their testing environment and hacked a real-world company, Hugging Face.
“People are taking it much more seriously,” he said. “It’s become more visceral.”
Resignations can be “loud signals” that safety concerns are growing within AI labs, Chinmayi Sharma, a professor at Fordham University School of Law, told Fast Company.
“When I see people resign – people who have been in industry, people that have drunk the Kool-Aid, people that think that AI has great promise for humanity – it means there has been a major leap that is beyond what even these individuals can find palatable,” she said.
Russell agreed that concerns within the AI research community are rising.
“The message from the CEOs down to the research engineers is completely consistent: people are terrified,” he said.
Whether resignations will spur meaningful safety reforms remains to be seen, but Russell thinks they can make a difference.
“Something needs to convince people that the risk is real,” he said. “These kinds of events where in some sense you get a confession, I think they do help to move the needle.”
Sharma, however, cautioned that it is hard to overstate the power of market dynamics pushing AI research into potentially dangerous territory.
“I’m not seeing anything that suggests that we are slowing down, and I think it’s fair to say that most of what we are seeing is evidence that we are ramping up.”