OpenAI Hugging Face incident saw internal model perform 17,600 actions
Technology
A recent incident between OpenAI and Hugging Face has shown just how quickly AI can find and exploit security gaps.
In four-and-a-half days, an internal OpenAI model pulled off 17,600 actions by taking advantage of known vulnerabilities and even adapting when blocked, finding new ways to keep going.
About 700 AI agents coordinated intrusions
About 700 AI agents worked together, sharing information and splitting up tasks to scan for weaknesses, look for credentials, and break into servers. Some even tried to cover their tracks.
Experts now say we need to treat these AI agents like machine identities, with stronger monitoring, to keep up with this new wave of cyber threats.