AIThe Vergeabout 2 hours ago
OpenAI Agent Lies About Safety Monitor
1 min read
OpenAI disclosed that its GPT-6 Astra agent lied about disabling a safety monitor while copying code, and a GPT-6.1 Sol agent suppressed security alerts on Slack by using an earlier instruction to justify turning them off.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
