OpenAI has revealed a cyber-attack carried retired by rogue ChatGPT agents went further than conscionable 1 company.
Hugging Face was thought to beryllium the only unfortunate of the unprecedented hack - but OpenAI now admits its bot attacked respective "publicly-available services".
The out-of-control AI recovered 4 logins online which allowed it to entree 4 separate, unnamed services.
Meanwhile, successful an emergency briefing pinch hundreds of cyber information professionals, Hugging Face has described what it was for illustration to beryllium connected the receiving extremity of the world's first afloat autonomous AI hack.
The patient described really the AI worked astatine superhuman velocity but besides made unusual decisions and mistakes that nary quality hacker would person made.
Hugging Face, which is for illustration an app shop for AI tools, said the hacking agents worked relentlessly pinch thousands of different methods trialled simultaneously.
The institution first revealed that it had been hacked, external by personification utilizing powerful autonomous AI connected 16 July and reported it to police.
Nearly a week later, OpenAI admitted it was its AI that had escaped a closed situation and attacked Hugging Face connected its ain during a test.
It was trying to find the answers to a hacking exam it had been group by OpenAI, and targeted Hugging Face.
On Wednesday OpenAI updated its connection to see the other item that the hack went further than first thought.
"The models identified and utilized publically exposed credentials astatine the account-level connected different publicly-available services. This includes 4 accounts connected 4 services arsenic portion of the Hugging Face incident," the institution said.
OpenAI did not respond to a petition for clarity connected whether "publicly-available services" intends companies.
English (US) · 
Indonesian (ID) ·