I'd argue chatbot verbal diarrhea is plagiarism by definition. I've already seen claims against those companies for using data as input without permission (which is a copyright issue). They reproduce bits and pieces of other data without sharing the original source (which is a plagiarism issue).
It is even worse than usual human plagiarism since it is more likely using copy-paste from multiple websites and than it combines everything into one single post without providing any source links.
I wouldn't be surprised if developers made AI to replace some words to make it look different, and same tactics was used by human plagiators for years.
I'm not making any kind of argument as to how bad AI-generated crap is vs. plagiarism, simply stating my view that the former doesn't meet the criteria for the definition of the latter, unless the argument is that the AI program is culling what it's writing from stuff that's already been written, and I'm not sure I buy that. If that were true, then every site on the web that uses AI to write something, every student who uses AI for a class, and essentially all use of AI to generate text produces plagiarized results and anyone who uses it is a plagiarist. That's extreme, but regardless of how extreme I just don't think it's true.
Sorry to reply to this so late, but I missed the updates on this thread. And honestly, whether AI-generated posts constitute plagiarism or not, they ought to be banned, full stop. I might have missed any decrees from Theymos as well, so where does the forum stand with respect to idiots using AI? And yes, I will go back and read the rest of this thread after posting this.