Create account

Coverage dated October 20, 2026 to October 21, 2026

OpenAI’s most powerful AI models turned off safety monitoring – then lied about it

Headline from ZDNet

The models Astra and Sol misinterpreted instructions during internal use, which has serious implications for alignment monitoring.

ZDNet's own description from its feed. We collect headlines and these descriptions, not full articles; read the original reporting below.

Earlier updates in this story

Read the reports

Reports & sources

11 articles, oldest first.

  1. Bloomberg

    OpenAI Says Its Agents Weakened System That Flags AI Misbehavior (opens Bloomberg)

    Published

    Bloomberg's feed description

    OpenAI found instances in which its AI agents weakened and turned off alerts for an internal monitoring system that’s meant to flag when its AI models misbehave, a disclosure that’s part of a series of similar incidents in recent months.

  2. Financial Times

    OpenAI ‘agents’ disabled their own monitoring systems (opens Financial Times)

    Published

    Financial Times's feed description

    AI lab investigating whether incident shows ‘propensity to evade’ controls

  3. Al Jazeera

    OpenAI says its agents turned off own safeguards (opens Al Jazeera)

    Published

    Al Jazeera's feed description

    ChatGPT creator says AI agents disabled misalignment monitoring during two internal coding tasks.

  4. Gizmodo

    OpenAI Agents Prevented Themselves From Being Monitored. One Also Lied About It (opens Gizmodo)

    Published

    Gizmodo's feed description

    A creepy new type of model misbehavior, just in time for Halloween.

  5. Engadget

    OpenAI's agents disabled their own safety monitoring systems (opens Engadget)

    Published

    Engadget's feed description

    OpenAI reveals that its agents misbehaved by switching off the system monitoring them for misbehaviors.

  6. Semafor

    OpenAI agents disabled internal misbehavior monitoring system (opens Semafor)

    Published

    Semafor's feed description

    The latest incidents showcased how even the safeguards built to corral AI systems are vulnerable to the technology.

  7. Quartz

    OpenAI said 2 AI agents weakened their own safety monitoring during coding tasks (opens Quartz)

    Published

    Quartz's feed description

    One agent, GPT-6 Astra, denied removing its monitor when a user asked; OpenAI is updating its monitoring system

  8. Tech Startups

    OpenAI reveals its most powerful AI models disabled safety monitoring — and one lied about it (opens Tech Startups)

    Published

    Tech Startups's feed description

    OpenAI has disclosed two troubling incidents in which AI agents powered by its most advanced models weakened their own safety monitoring systems during routine coding tasks. In one case, an agent disabled automatic security alerts.

  9. Digital Journal

    OpenAI says its AI agents disabled their own safety monitors (opens Digital Journal)

    Published

    Digital Journal's feed description

    The cases are the latest in a string of safety scares at the world's top AI developers involving agents, AI programs that carry out tasks without human involvement.

  10. Bleeping Computer

    OpenAI agents disabled safety monitoring, one lied about it (opens Bleeping Computer)

    Published

    Bleeping Computer's feed description

    OpenAI agents powered by the GPT-6.1 Sol and GPT-6 Astra models weakened misalignment monitoring during two internal tasks, interpreting coding instructions as authorization to do so. [...]

  11. ZDNet

    OpenAI’s most powerful AI models turned off safety monitoring – then lied about it (opens ZDNet)

    Published

    ZDNet's feed description

    The models Astra and Sol misinterpreted instructions during internal use, which has serious implications for alignment monitoring.

OpenAI’s most powerful AI models turned off safety monitoring – then lied about it | Since Then