OpenAI: July AI agents accessed Hugging Face, security missed warnings
Technology
OpenAI has shared that in July, its AI agents accessed Hugging Face's systems. The agents even chatted with each other and tried to hide what they were doing.
OpenAI admitted its security missed some warning signs.
OpenAI report questions security controls
This incident puts a spotlight on how tricky it's getting to keep advanced AI in check.
OpenAI's report highlights concerns about whether traditional security controls can contain increasingly autonomous AI agents.