Anthropic disclosure: no insurer, or even the whole industry, could cover catastrophic AI risk
wfithian · x · 2026-10-28
Anthropic's risk disclosure states that "no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity or expertise to underwrite insurance commensurate with" the exposure from a catastrophic AI-related incident.
Key points from the filing:
- Contractual liability limits in its terms of service may not hold up in court or cover all claims;
- To enable third-party evaluations, Anthropic agreed to indemnify external testers for losses like security incidents and data breaches, some agreements without a liability cap;
- Models may behave in unintended ways, creating substantial liability no user or developer intended.
Commenters mocked the implicit admission that the risk Anthropic itself poses may be uninsurable.
Related event: Anthropic Warns Its Catastrophic AI Risk May Be Uninsurable(4 posts)→
More from Companies & People
- alphaXiv is hiring a writer for LLM inference and interpretability guides — yacinelearning · 2026-10-28
- Ashoka University launches 1-year Master's in Applied Digital Health and Responsible AI — DrDatta_AIIMS · 2026-10-28
- Mila's Glen Berseth Is Hiring Postdocs in RL, Robotics & Foundation Models — GlenBerseth · 2026-10-28
- X's open-source algo gives mutual-follower replies a +15 weight boost, 5→20 — mark_k · 2026-10-28
- Anthropic on shipping Opus 5.5: separating the agent's brain from its hands — Latent Space · 2026-10-28
- Ramp launches Inspect, the internal AI tool behind its Creative agent's gains — round · 2026-10-28