I saw that chatGPT retired their AI detector recently, because it was not accurate.

As you noted, it's common to get wrong results from the AI detection sites. The best option I've come up with is to use multiple sites and check multiple posts by the same author. Even after that, I am rarely 100% confident.

I agree with you about the people who make strong accusations in comments. I don't know what the right answer is, but aggressively accusatory comments are usually not helpful - IMO.

It's a hard problem, but we definitely don't want false accusations to drive away good contributors.