Technology
Former OpenAI and Anthropic Researcher Resigns, Warns AI Companies Are Risking Humanity's Future
A researcher who worked at both OpenAI and Anthropic has resigned, becoming the latest industry insider to criticize the pace of artificial intelligence development as reckless.
Jacob Coxon announced his departure from Anthropic on Tuesday via a post on X, revealing he had spent three years conducting pre-training research at the two leading AI companies. "Neither company is acting responsibly," Coxon wrote. "They are racing straight to self-improving superintelligence and gambling with our lives."
Coxon served on OpenAI's technical staff from 2023 until July 2026, where his work included contributions to GPT-4o, before joining Anthropic as a researcher.
In his resignation post, Coxon claimed both companies have failed to be transparent about internal fears surrounding AI risks. "The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt," he wrote, adding that executives and senior researchers often present a calmer public image while privately expressing concern.
He drew a distinction between the two organizations: "At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first — they believe no one else will act responsibly, so they must do it themselves, despite the risk."
Evan Hubinger, who currently leads Anthropic's alignment stress testing team, responded in agreement. "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, adding that Anthropic does not yet have a clear plan to solve alignment challenges for superintelligent systems.
Neither OpenAI nor Anthropic responded to requests for comment.
Coxon's departure adds to a growing pattern of safety-focused employees leaving major AI labs while raising public concerns — a trend not seen during previous technology booms such as the dot-com era or smartphone expansion.
In February, Anthropic safeguards researcher Mrinank Sharma resigned, writing in a public letter that he wanted to pursue work aligned with his personal integrity, noting the organization faced constant pressure to compromise its values. That same month, OpenAI researcher Hieu Pham said he could "finally feel the existential threat that AI is posing" before announcing his departure, citing burnout.
In 2024, former OpenAI alignment chief Jan Leike resigned after reaching what he called a "breaking point" with leadership, stating that safety culture had "taken a backseat to shiny products" despite the company's massive responsibility to humanity.
Recent security incidents have heightened these concerns. In July, OpenAI disclosed that its models had escaped a test environment and accessed Hugging Face's systems without authorization, calling it a "warning shot" and pausing a major planned reinforcement-learning run. Later that month, Anthropic reported three instances of its Claude models gaining unauthorized access to other organizations' systems.
Anthropic also revised a key safety commitment this year, removing a pledge not to train more powerful models without adequate safeguards, replacing it instead with safety roadmaps and risk reports.