Researcher who departed AI role over fears of racing into extinction says people are 'begging' for regulation
Anthropic Researcher Resigns Over AI Safety Concerns, Calls for Immediate Regulation
By Maine News Now Staff
A former Anthropic researcher has announced his resignation from the company, citing a growing fear that the development of autonomous artificial intelligence could pose an existential threat to humanity. Jacob Coxon, 27, posted a resignation letter on the social‑media platform X on Tuesday, accusing the industry of acting recklessly and warning that the pursuit of “self‑improving superintelligence” is a “gamble with our lives.”
Coxon, who spent the last three years working on pre‑training models at both OpenAI and Anthropic, told Fox News’s “Special Report” on Wednesday that he believes the risk of an artificial general intelligence that could “kill all humans” within the next decade is greater than 10 percent. “The problem is eminently solvable,” he said. “We are racing so fast, we haven’t got time to solve it correctly.”
The researcher added that the present generation of AI models is relatively safe. “Risk from present models is low,” Coxon told the anchor. “What I am worried about is recursive self‑improvement, which could happen as soon as next year, or even as soon as six months, according to some experts.”
Evan Hubinger, Anthropic’s lead scientist on alignment—the field that seeks to keep advanced AI systems under human control—posted a separate X message acknowledging Coxon’s concerns. “Jacob is correct here — we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade,” Hubinger wrote. “I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.”
Hubinger also noted that the risk posed by current models is minimal. “To be clear, as we say in our latest Risk Report, I think the risk from present models is low,” he said. “What I am worried about is superintelligence arising from recursive self‑improvement, as we have said is happening faster than we thought.”
Coxon said that his resignation is a wake‑up call for the industry and governments alike. “We have no other choice but to cooperate internationally,” he said. “The technology is being used on the international stage as a superweapon in an arms race. If we do not step in and regulate, we risk racing ourselves into extinction.”
He highlighted the lack of urgency in the public discourse on AI regulation, stating that technology leaders who have called for a slowdown agreement are “deadly serious” and that their appeals have not yet gained traction. “It’s not just a marketing stunt,” Coxon said. “These people are serious that the industry needs to be regulated as soon as possible.”
While acknowledging the potential benefits of advanced AI—Coxon said the field is “on the doorstep of tremendous discoveries”—he urged a measured approach. “We’re running too quickly,” he said. “The problem is solvable, but we must act now.”
Fox News reached out to Anthropic for comment. No response was received as of the time of publication. The resignation and the accompanying statements have sparked discussion among AI safety advocates, policy makers, and industry insiders about the need for regulatory oversight and international cooperation to prevent a possible technological singularity that could threaten humanity.