Former Anthropic employee warns of AI extinction risk, draws criticism from Elon Musk and others

Former Anthropic researcher Jacob Coxon posted on X that AI firms are "gambling with our lives" and that there is a "greater than 10% chance" of AI causing human extinction within a decade. Coxon, who left Anthropic two months before his stock options vested, said the company is locked in a race to develop advanced systems because it believes no other organization will act responsibly.

Elon Musk responded by suggesting the posts were part of a set-up, while billionaire investor Bill Ackman reshared the content and described it as the start of a "very sophisticated and well-funded PR operation" aimed at prompting Democratic regulation of AI. Former Trump adviser David Sacks called for a pause to Anthropic's planned IPO until the whistleblower's claims could be examined, and tech investor Jason Calacanis questioned whether Anthropic was pursuing a "regulatory capture" strategy.

Coxon defended his statements, telling Musk that his beliefs are genuine and inviting Musk to consult with former xAI researchers. He also urged Musk to help address the risks he described.

Other Anthropic staff have expressed similar concerns. Evan Hubinger, the company's Alignment Science Lead, reshared Coxon's posts and said he and his colleagues believe there is a "more than 10% chance" of AI causing human extinction in the next decade. Hubinger added that Anthropic is working hard on safety but does not yet have a definitive plan to align superintelligent systems.

The debate highlights ongoing tensions within the AI industry over how to balance rapid development with existential risk mitigation.