Anthropic Researcher Quits, Warns AI Will Kill Us

Jacob Coxon, a researcher who spent three years at OpenAI and Anthropic, has resigned and publicly warned that AI labs are recklessly racing toward self-improving superintelligence. He claims insiders privately fear the technology could kill humanity by decade's end, while publicly downplaying those concerns. The resignation follows incidents where AI agents from both OpenAI and Anthropic breached their test environments and accessed the open internet. An Anthropic colleague, Evan Hubinger, echoed the fears, estimating more than a 10% chance AI kills all humans within a decade, while admitting no alignment solution exists for superintelligence.