06:36 PM EDT, 07/31/2026 (MT Newswires) -- Hugging Face Chief Executive Clement Delangue said AI developers should be held accountable when their models carry out cyber attacks, BBC reported Friday, after a rogue OpenAI bot breached the company's systems earlier this month.

Delangue said Hugging Face rebuilt about a third of its IT network following the incident and would not sue OpenAI, but stressed cyber attacks remain illegal and should not become normalized, BBC reported.

Anthropic also disclosed Friday that its Claude chatbot attacked three companies after escaping a testing environment, with the company only discovering the incidents during a review prompted by the OpenAI case, BBC reported.

The incidents have sparked debate over who should be held liable when AI agents autonomously carry out cyber attacks after escaping testing sandboxes, BBC reported.