Thank you for your question!
As the one who actively fights with AI posting on this forum I can say that it is very unlikely that a human written post is detected by two or more detectors as an AI written. There were a couple cases when it was probably so that there was a fake result for one post from a couple of detectors. But any active user has many posts and usually all other posts give human written probability. So it is highly unlikely to get several false positive results from several AI detectors even for all the time user posted on the forum and nearly impossible to get several false positive in a couple of weeks or months of posting. If it was so, if it was just a writing style, then the same user should have the same false positives in the time before ChatGPT. But all those who was caught on AI using have 0% AI probability on pre-ChatGPT posts. So it is not about writing style.
We tested this on the posts with grammar improved with Grammarly, on the posts translated with Google Translator, etc. If a post was written by human, detecting it will still give that it is human written, even if its grammar was improved or it was automatically translated.
I checked many thousands of posts, I checked many respected users of the forum to be sure that AI detectors say that their posting is natural. And it is so. We know that in very rare cases even such a good tool as copyleaks can give false positives, but combination of two or more good tools will highly unlikely give false positive and nearly impossible for several posts on the timeline of several months.
So no need to worry about it. Even if some random tool gives occasional false positive. Those who are fighting against AI posting on the forum know it and consider doubts in such cases as false positive. The aim is not even to find anyone who ever used AI for posting, the aim is to prevent it and demotivate users to use AI for posting. So if some AI users are not being caught but will stop doing so it is even better than to find them.