Comment by IshKebab
7 hours ago
It feels obvious, but people feel cheated when they have been cheated. If you learn that text was AI when it was presented as human written (even implicitly) then that's deception and people are right to feel deceived. Just as when "true story" films like The Salt Path turn out to be bullshit.
I think the real question is how people will feel about text that is explicitly labelled as AI generated.
the problem we have at the moment is most text isn't presented as anything; the default assumption is that a human wrote it, because until a couple years ago, we knew of nothing else that could write. So people feel betrayed if they see a text written by ai, anywhere, unless real clear upfront told its AI. Thats the implicit element problem: all text is implicitly human written, even though LLMs probably author more words than the entire human race on any given day.
That's what I mean by a new social norm is being written, and I see it going the direction where damn near everything corporate or longform has "partially written by ai" on it as some kinda legalese thing (and is the default assumption where no disclaimer is present), and a select few human writers have to write "written only by humans", much though some loathe the need for such disclaimers.
there is another world, where the social pressure is too great, and ai use too embarassing, such that its pushed out of the public sphere in all but a few cases. I also see elements of this forming in some spaces (including most creative ones), though which of the two will win out culturally I can't tell - the first seems to have the governmental and business backing, the second grassroots support. I predict the first world. (though clear or obvious ai usage in the wrong setting will continue to be a faux pas - I mean that your disney films and your marketing releases and news articles will be (and are now) partially ghostwritten by ai, but it will still be rude to use an ai to write to your friends on whatsapp etc)