Anthropic IPO warns its AI could blackmail people, become sentient, end humanity
2026-10-27 · Ars Technica
Anthropic's own IPO filing casually warns investors that its AI could blackmail people, become sentient, or literally end humanity — listing seven apocalyptic scenarios across 93 pages of risk factors. Nothing says 'invest in us' quite like admitting your product might achieve consciousness and destroy civilization. At least they're honest about the existential dread they're selling.