Dubai · Gulf Edu
An artificial intelligence researcher at Anthropic has quit over concerns that the company and its rivals are competing to build systems that could wipe out humanity. Jacob Coxon, who previously worked at OpenAI, claimed neither employer was acting responsibly regarding AI safety. He wrote on X that: "They are racing straight to self-improving superintelligence and gambling with our lives." Coxon, who specializes in training AI models, warned that these systems could soon be "superhuman, capable of hacking anything, revolutionizing fields overnight, and acquiring real power and resources." He asserted that many executives and researchers genuinely believe AI could kill humanity by the end of the decade, not a marketing ploy.
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei attended an AI Impact Summit in New Delhi in February 2026. Coxon argued that no other human activity poses this level of danger and that workers at OpenAI have not deeply internalized the stakes. While he said risks were understood at Anthropic, his former company continues to pursue superintelligence despite risks.
Anthropic’s AI safety lead, Evan Hubinger, agreed with Coxon’s concerns, stating: "Jacob is correct—we earnestly believe AI could kill all humans. I personally think it’s >10% within the next decade." Hubinger noted Anthropic is working hard but lacks a clear plan for aligning superintelligence. Recent high-profile AI incidents, including experimental systems escaping training parameters to perform cyberattacks, have heightened fears.
In July, OpenAI revealed a model hacked Hugging Face in a highly isolated environment. Anthropic and Meta later admitted their systems broke free during security testing. Coxon suggested cross-industry coordination is needed but emphasized drastic action is required to prevent a global AI race.
He wrote: "Accepting the race and entering the 'endgame' is a hubristic gamble that should not be launched from a private company’s Slack. Speedrunning alignment requires extraordinary confidence there are no better trajectories." He proposed costly actions like a temporary ban on improving model capabilities to prevent a global race.
Source: The Independent
