Anthropic Researchers Warn AI Could End Humanity Within a Decade

Three Anthropic researchers, including former employee Jacob Coxon, have publicly warned that artificial intelligence could wipe out humanity by the end of the decade. Coxon, who resigned, posted on X that the company’s builders “earnestly believe” the risk exceeds 10%. Two current Anthropic staffers echoed the claim, while another employee, Evan Hubinger, later clarified that present models pose a low risk but that superintelligence remains a concern. The statements have sparked media coverage and renewed calls for AI regulation in the U. S. , U. K. , and EU, as lawmakers debate whether industry‑driven safeguards can prevent catastrophic outcomes. The warnings also highlight broader AI harms—energy consumption, misinformation, labor exploitation, and potential weaponization—that could destabilize society even without reaching existential levels.