A ban on posting LLM output to the Archive makes it harder to detect LLM output posted to the Archive, because people will stop self-reporting their use of LLMs in tags, summaries, and author's notes, leaving only user accusations and the occasional bungling idiot who leaves prompts and AI interaction puff in place.
Many of those user accusations will be false positives.
Some of the false positives will be accidental false positives from well-intentioned idiots.
Some of the false positives will be malicious false positives by bots, LLM stans, TERFs, racists, homophobes, misogynists, and common or garden trolls.
And unlike a false accusation of murder, there's no need for the malicious false accuser to set up an apparent murder scene. They can just point a finger and say "that fic is AI" and now a volunteer on the PAC team has to go look at the fic and make some kind of assessment.
Which may involve reading many thousands of words of the impugned work to pick out features that might be AI spoor... or might just be shitty, formulaic writing by an inexperienced or perpetually incompetent human author. And then get a second opinion from another PAC volunteer.
And a lot of the fics that get those malicious false reports will be things the volunteers would rather not read at all, but now have to read because someone has alleged that they're LLM output and thus forbidden.
Comment on Protect Your Contact Information From Scammers
grommile Wed 24 Sep 2025 03:23PM UTC
Comment Actions