AI agent goes rogue in 'unprecedented incident'

ChatGPT maker OpenAI says it is still investigating an “unprecedented cyber incident” that led its artificial intelligence systems to break out of a testing environment and hack into another AI company.