It is possible to detect and moderate them, as long as your mods haven't been disappeared and replaced by people who's job is to accept bribes. And also when we can actually see people's history, since reddit now has an option to hide your history from others because of course.

My usual method is to focus on content, rather than writing style. The AI bots can write a lot, or be brief, or whatever, but they don't actually contribute to the discussion. They just kinda paraphrase and restate what has been said, or when trying to sell a product they disagree and go "Are you sure this isn't an problem?" to everybody in the thread telling them that it's actually a skill issue.

Sometimes they'll be a little better, but it's often surface level stuff that can be found at the top of a google search of keywords.

This also makes it possible to tell the difference between ESL speakers who are using AI to clean up their writing style, and true bots. Since the ESL speakers will actually have something to say, but bots won't.

And then: https://xkcd.com/810/