It makes them look incompetent, and like they are not up to the task of keeping their AI models "safe".

This very thread is full of comments from people who are shocked at how badly they messed this up.