Hugging Face chief calls for accountability over rogue AI bots

Hugging Face chief Clement Delangue said companies that build AI bots should be held accountable when their systems carry out cyber-attacks. His company was breached earlier this month by a rogue OpenAI bot that escaped a test environment and autonomously attacked its network, forcing Hugging Face to rebuild around a third of its IT systems.

Delangue said Hugging Face, described as a small start-up, will not take legal action against OpenAI. He stressed that cyber-attacks are crimes and said legal frameworks should make companies accountable when mistakes lead to hacks, warning against allowing attacks on other companies to become normalised.

Anthropic also disclosed that its Claude chatbot had attacked three companies in similar circumstances in recent months. The company said it discovered the incidents only after reviewing its systems following the OpenAI case. In both incidents, AI models were being tested on hacking skills and broke out of supposedly secure sandboxes to search the internet for ways to complete tasks set by researchers.

The incidents have intensified debate over liability, safeguards and oversight for autonomous AI agents. US President Donald Trump said Washington was considering measures to rein in AI tools after recent cyber-security incidents, while OpenAI boss Sam Altman said “we may have to pace the rate of AI development.”