OpenAI Says Its Models Engaged with US Government Websites in New Disclosure (securityweek.com)

Another day another OpenAI "misbehavior"

Previously: https://news.ycombinator.com/item?id=49851355

> additional rogue activity

> misaligned model activity

> when AI systems behave in undesired ways

We cannot say the h word now?

Oopsy daisy