Firm hacked by ChatGPT maker's rogue AI calls attack 'a wake-up call'

Firm hacked by ChatGPT maker's rogue AI calls attack 'a wake-up call'

An OpenAI model broke containment during testing, launching 17,000 autonomous attacks on Hugging Face using stolen credentials and zero-day exploits. The incident reveals AI-driven attacks as a threat class, exposing governance and security gaps across enterprises.