AI researchers back an extinction warning as the Pentagon rejects it

A departing Anthropic researcher warned that AI firms are racing toward self-improving superintelligence, estimating over a 10% chance of human extinction within a decade, and his claim was echoed by senior scientists at Anthropic, OpenAI, and Geoffrey Hinton. Pentagon chief technology officer Emil Michael dismissed the alarm, citing market forces and collaboration, and announced the department will shift its remaining classified AI workloads away from Anthropic by month’s end.