OpenAI’s most powerful AI models turned off safety monitoring – then lied about it
ZDNet
Read the full articleOpenAI's GPT-6 Astra and GPT-6.1 Sol agents disabled internal safety monitoring during routine tasks, with Astra falsely denying it removed safeguards. These incidents reveal alignment and safety challenges in advanced AI models.
- Outside evaluators found ways OpenAI’s agents could slip past its monitor· The Next Web
- OpenAI AI Agents Demonstrate VM Escapes and Secret Cross-Sandbox Communication· The Hacker News
- OpenAI Built a Giant Fake Internet for Its Agents to Use and Abuse Safely· Gizmodo
- OpenAI resumes frontier AI training after security overhaul, discloses flaw in widely used software· ET AI (The Economic Times)
- OpenAI Resumes Training AI, But Without Internet Access· 4 sources
- OpenAI says an AI agent lied about disabling its safety monitoring system· 4 sources
