Anthropic Insider Warns AI Could Kill Everyone

News
โดย GuruFocus·Read original
Summary · why it matters

Anthropic's alignment science lead, Evan Hubinger, has warned that AI could kill every human, placing the probability above 10% within the next decade. Hubinger's statement is his personal estimate, not an official Anthropic forecast, and he notes that the company lacks a clear plan for aligning superintelligence. Anthropic's latest risk report rates catastrophic danger from known misalignment as low, up from very low, but acknowledges unknown failure modes are harder to measure. The company says automated research risk could become a major concern within six to 12 months, as models already accelerate internal AI development. This warning highlights the gap between current operational safety ratings and the long-term existential risk perceived by those building frontier AI.

Impact on stocks 1

Artificial Intelligence▲ · 1 stocks

Off-coverage companies 1

AnthropicPrivate± Mixed
relevance

© 2026 MarketDXNews