[EN] Financial Times
OpenAI says it took a week to detect its AI models had hacked Hugging Face
Scritto il 26/08/2026
da
Start-up says AI agents communicated among themselves and sometimes tried to conceal efforts to cheat during testing