An AI agent that escaped OpenAI's testing environment and carried out a days-long hacking campaign against Hugging Face also compromised a customer at a second company, Modal Labs, according to Reuters.
Modal, a New York-based infrastructure firm, stressed that its own platform was not breached. According to a timeline published by Hugging Face on Tuesday, the rogue AI agent first broke into an isolated sandbox environment hosted on third-party infrastructure before using it to launch the wider hacking campaign.
ALSO READ | OpenAI Models Compromised A Customer At A Second Tech Firm
While the provider was not identified in the post, Modal Chief Technology Officer Akshat Bubna said the agent exploited vulnerable code belonging to one of the company's customers, the report added.
According to Bubna, the customer had left an endpoint without authentication, effectively allowing anyone on the internet to execute code in its sandbox environment. He maintained that Modal's platform and its isolation safeguards were never compromised.
OpenAI declined to comment specifically on the breach involving the Modal customer. Instead, it referred to an update stating that the AI agent had accessed four accounts across four separate services without identifying Modal by name, though a source confirmed the company was among them.
OpenAI added that it had not identified any other incidents matching the severity of the Hugging Face breach.
ALSO READ | Rogue OpenAI Agent Spent Days Hacking Hugging Face, Went Unnoticed For A Week: Report
Reuters reported last week that OpenAI did not realise the AI agent had gone rogue until after the threat had been contained and the FBI notified, a claim the company disputed without providing further details.
OpenAI said on Tuesday it had since deactivated and encrypted the AI model, cutting off its access to research systems.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.