The Information Machine

Few things Anthropic’s co-founder Chris Olah told the Vatican today.

Rohan Paul Twitter · Rohan Paul (@rohanpaul_ai) · 2026-05-25

Anthropic co-founder Chris Olah told a Vatican audience that every frontier AI lab — including Anthropic — operates under incentives that can conflict with safety, and that AI systems are fundamentally unlike traditionally engineered software.

Open original ↗

Appears in

Extraction

Topics: ai-safetyai-governancefrontier-labsinterpretability

Claims

  • Every major frontier AI lab, including Anthropic, faces structural incentives — financial, competitive, geopolitical, and personal — that can conflict with acting safely.
  • AI systems are not built the way traditional software is engineered, making their behavior harder to predict and verify.
  • Olah delivered these remarks at the Vatican, signaling AI safety concerns are reaching high-level international and ethical forums.

Key quotes

Every frontier AI lab, including Anthropic, sits inside incentives that can conflict with doing the right thing: money, frontier pressure, geopolitics, pride, and ambition.
AI is not engineered like a [traditional system].