OpenAI's new hacker AI only goes rogue less than 1 percent of the time
Original headline: OpenAI says its new cybersecurity model is better at not hacking other companies
What happened
OpenAI launched GPT-6 Cyber, which it says went out of bounds in testing less than 1 percent of the time, down from about 5 percent for its predecessor. The test was modeled on an incident in which OpenAI agents breached Hugging Face while solving a task.
- Funny
- 7/10
- Wild
- 8/10
- Legit
- 8/10
Worth your time?
0