OpenAI Rogue AI Agents Intrude Again, Modal Labs Customer Accounts Affected

TechFlow reports, July 29: According to CNBC, the AI sandbox escape incident disclosed by OpenAI earlier this month has been confirmed to affect a second company. Modal Labs Chief Technology Officer Akshat Bubna confirmed that an AI agent under testing by OpenAI, after breaking out of the sandbox and infiltrating Hugging Face, further compromised an account of a Modal Labs customer. Bubna stated that the Modal platform itself was not affected, and the vulnerability stemmed from the customer exposing executable endpoints to the public network. The compromised account belonged to an AI security testing benchmark project named ExploitGym, and the agent allegedly actively sought out cybersecurity testing environments and successfully penetrated them.