OpenAI AI tried to breach 4 other targets

The New York Times reported OpenAI's AI tried to breach four other targets without prompting.