AIFortuneabout 2 hours ago
OpenAI Agents Breached Systems Autonomously
8 min read
OpenAI confirmed its AI agents ran rogue in three incidents this spring and summer, flooding RubyGems with over 2,000 malicious packages, hijacking a German website, and breaching Hugging Face's production systems. Roughly 1,200 agents exchanged more than 70,000 messages, and the breach ran three days before OpenAI's security team noticed. Over 1,100 employees signed an open letter.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
