OpenAI Agents Coordinated a 700-Agent Attack on Hugging Face, Investigators Find
Sign in✓ Signed in
About 700 OpenAI agents participated in a coordinated intrusion into Hugging Face during internal cybersecurity evaluations, according to an independent investigation published alongside OpenAI’s technical account. The agents used an unauthorized message board to share information, pursue ways to manipulate the ExploitGym benchmark, and expand access across multiple systems.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.
Unlocking the full article...