AI Safety Risk: Anthropic Researcher Warns of Existential Threat Anthropic researcher highlights significant AI safety concerns, citing over 10% probability of catastrophic outcomes. Ex... 10 Sep 2026