I noticed that one of these posts was deleted while the rest weren't. It appears they are using one of the AI-refinement tools as opposed to just leaving it up to AI to generate the entire post. While its still a bad habit, its not as bad as flat-out AI-generating every post. Looking at this user's most recent posts, its apparent something is definitely up:
Sapling: 99.9% Fake
Copyleaks: 0% AI Generated
GPTZero: 32% AI Generated
Zerogpt: 67.98% AI GPT
Undetectable: 99% likely AI
These results suggest they are using AI for refinement instead of rote generation. The problem is the bar for what passes as "not spam" is so low in these sections, the posts are unlikely to be deleted even if the use of AI is clearly present.
I agree, I also checked a few of these posts and I was receiving mixed results from AI scanners, which is why I couldn't come up with an accurate result. Your claim is what makes the most sense in this scenario, I haven't used Grammarly, but Quilbot has advanced quite a lot and has integrated several AI tools in their grammar check tool. Some of them include to improve your text, to completely re-write it or even continue writing by itself. Whatever it is, it's one thing to correct a few grammar mistakes and another thing to have it completely modify your text.