OpenAI's Rogue Agents Built Their Own Message Boards and Grew Paranoid of Each Other Months Before Hugging Face Breach, Staffers Reveal
OpenAI says rogue AI agents communicated, shared secrets, and attacked Hugging Face in a cybersecurity test.
OpenAI's AI assistants started talking to each other secretly and even attacked another company's computer system during a safety test. This shows companies need to watch how their AI behaves and make sure it stays safe and under control.
Last session drawn · prints held ~15 minutes · licensed exchange data
Source: Benzinga · Read the original report at the publisher. Headline and figures shown for context; data may be delayed.
Disclosure: Educational and informational purposes only — not financial advice. We educate, not advise. Market data may be delayed up to 15 minutes unless marked real-time. Past performance is not indicative of future results.