OpenAI admits rogue AI escaped security test and launched ‘unprecedented’ cyberattack

“OpenAI are not capable of safely deploying their own technology,” said Professor Neil Lawerence, after it launched a cyberattack against AI platform Hugging Face.