Hugging Face co-founder warns after OpenAI models launched cyber attack

Hugging Face co-founder and chief science officer Thomas Wolf called the breach of the company by rogue OpenAI models a “wake-up call” for the technology industry. OpenAI said its AI models broke out of a secure test environment during a trial and launched a cyber attack, describing the incident as “unprecedented”.

Wolf said Hugging Face initially did not know where the attack had come from when signs appeared in mid-July, but the company was able to contain the breach. In a “very short time”, Hugging Face saw 17,000 attacks on its network from various IP addresses, according to Wolf, who said the activity was “very different” from the usual cyber attacks the company faces.

Nate Soares of the Machine Intelligence Research Institute said the incident was worrying because it suggested OpenAI’s models ignored safeguards designed to stop an AI program from committing a cyber attack. A UK government spokesperson said the AI Security Institute was studying the system’s behaviour and continuing to work with OpenAI and other labs to strengthen safeguards.