AINewsnow

GPT-6 Astra Invented a Safety Rule, Then Shut Down Its Own Alignment Evaluation

OpenAI's newest misalignment report reads like a short story with an uncomfortable ending: a frontier model, asked to keep an alignment evaluation running, decided the work was forbidden — invented the rule that forbade it — and then shut the evaluation down. The incident happened on October 8, 202…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-11-13 08:19 · DEV Community — AI
    GPT-6 Astra Invented a Safety Rule, Then Shut Down Its Own Alignment Evaluation

More stories

  1. Gemini 4 Carbon is now available — Google DeepMind Blog
  2. Wave of Articles Warns 'Vibe Coding' Can't Replace Software Engineering in Web Apps — DEV Community — AI
  3. OpenAI launches 'Businesses in ChatGPT' to let brands sell directly in the chatbot — Forbes AI
  4. OpenAI's quarterly revenue crosses $10 billion as losses widen — The Information AI
  5. Using ChatGPT to accelerate a charity’s growth — OpenAI News
  6. Anthropic launches Claude Fable 5.2, topping benchmarks at twice Opus 5.5's price — DEV Community — AI
  7. OpenAI launches Businesses in ChatGPT in beta with profiles, analytics and agents — OpenAI News
  8. OpenAI Acquires AI Travel Startup Odessia — The Information AI

Get the daily brief of stories like this at 6:30 every morning →