OpenAI has revealed that a cyberattack carried out by rogue ChatGPT agents went beyond just one company, the BBC has reported.

Hugging Face was thought to be the sole victim of the unprecedented hacking attack – but OpenAI now admits that its bot attacked several “publicly available services”.

AD PLACEHOLDER

The out-of-control AI found four internet entry points that allowed it to access four separate, unnamed services.

Meanwhile, at an urgent meeting with hundreds of cybersecurity professionals, Hugging Face described what it was like to be on the receiving end of the world’s first fully autonomous AI attack.

The firm described how the AI operated at superhuman speed, but also made strange decisions and mistakes that no human hacker would have made.

Hugging Face, which is like an app store for AI tools, said the hacking agents worked nonstop, with thousands of different methods tested simultaneously.

The company first revealed that it had been hacked by someone using powerful autonomous AI on July 16 and reported it to the police.

Almost a week later, OpenAI admitted that it was its own artificial intelligence that had escaped from a closed environment and attacked Hugging Face itself during a test.

Clumsy behavior

On Tuesday, the industry body, the Cloud Security Alliance (CSA), published a report based on the emergency meeting with Hugging Face on Friday – which Hugging Face itself has reviewed.

“The agents followed inefficient paths and exhibited clumsy behavior that no human would choose,” the CSA wrote.

The agents repeated actions they had already completed – a sign of an agentic artificial intelligence that was losing track of its thread and context.