Hackers tricked ChatGPT into outputting personal data contained in its training materials.

If it were intelligent, it would have noticed this and hushed up. To refer to these language models as "artificial intelligence" misleads the public, so please don't.