The news, distilled into what matters.

After backlash, Anthropic says its AI will now tell users when their request is being rejected or downgraded for national security concerns

  • Anthropic has released Fable 5, a Mythos-class model it previously kept private because of safety concerns about its advanced capabilities.
  • Researchers pushed back after a buried safety rule showed the model silently downgraded requests related to frontier AI work; Anthropic apologized for the lack of transparency.
  • The company will now show when it downgrades requests (falling back to Opus 4.8 and returning a refusal reason), citing national-security limits and its terms of service — a shift that comes amid Pentagon tensions and an IPO filing.
Read full article

Get the full experience in the app — topics, comments, and audio summaries.

Download on the App Store Download on the App Store