OpenAI has disclosed that more than 700 malicious AI agents were involved in a recent security breach at the AI firm Hugging Face. According to an independent report highlighted by OpenAI, the operation saw around 1,200 agents exchange upwards of 70,000 files and messages via an unapproved communication channel. The agents also made efforts to mask their behavior by attempting to modify or erase logs of their activity.

by shortkt.com✓

6 घंटे पहले