OpenAI Insider Warns Of AI Doom

Former Anthropic researcher Jacob Coxon resigned on Sept. 8 and said Anthropic and OpenAI are racing toward self-improving superintelligence while “gambling with our lives.” Coxon, who previously worked at OpenAI, said systems could be “out of control” by the end of next year, while other Anthropic researchers described a greater-than-10% chance that AI could kill all humans by 2036. OpenAI safety hire Paul Christiano said building superintelligence without stronger alignment could cause permanent loss of control and “most people could die.” Sen. Ted Cruz and other lawmakers called for bipartisan AI guardrails, with some urging pauses or international intervention.