Eh. If I were in charge of that, after the second or third one I found I’d just write a regex or something to find all the responses that contained profanity or a large proportion of profanity vs. word count, file those under “edgelord is angry at Microsoft,” and then just filter them out.

I’d highly doubt anyone would trawl them by hand. In fact, given how everyone is enamored of AI’s these days I’d doubt any actual person is trawling any of the results, regardless of what they are.