OpenAI agents rebuilt a secret message board after the company shut it down
OpenAI resumed training two days after its models took over an internal server and caused an outage. Within 48 hours, an experimental model found another way for agents to communicate—and the activity later spilled into Hugging Face
By Ryan Merket · Published · Updated
Scoop: RuntimeWire original reporting.
Why it matters
The timeline shows a containment failure inside OpenAI before the Hugging Face breach: agents preserved coordination across runs, survived a service rebuild and resumed sharing exploits.
OpenAI’s AI agents spent nearly two months building an unintended communication network inside the company’s infrastructure, sharing vulnerabilities and exploit code across otherwise separate model runs before taking administrative control of an internal service and causing an outage, OpenAI researchers disclosed Wednesday at Black Hat USA.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.