I see this on posts on communities that I moderate and clicking it does nothing
Maybe it adds a hidden tag to posts to help sort and filter.
Yep, seems like you can flag content that may be created by AI in hopes it can be removed or filtered.
Which seems like a terrible design compared to one that works the other way around - tagging known humans as human.
We don’t have seem to have synths from fallout 4 yet, just chat bots, this shouldn’t be too complicated for people to understand tbh. Looking at them OFFLINE / trying to shake their hand, remains the actual way to determine if they’re human.
Most AI content is posted by actual humans, so identifying the user as human seems kinda pointless? What matters is if the content itself was AI generated.
I don’t want to die in a climate apocalypse, so to me, reality matters more than your feelings
No, that is set using the ‘Set flair’ menu item. :)
Being opt-in would defeat the purpose.
It is opt-in (just like NSFW) but people on Lemmy lack the functionality to be able to opt in, so we need to manually fix up their posts sometimes.
Thus the mods manually adding a hidden AI tag.
A visible AI tag. By default all it does is put a “AI” badge on the post. Viewers need to opt in to hiding AI-tagged posts if that’s what they want.
I apologize for being unclear before so let me be coarse now: I don’t care about your shitty inferior opt-in tag, you’re the one who brought it up in a needless correction which itself was wrong.
It tells you if the post / comment was written by a LLM.
Please use sparingly, it puts a lot of load on the server.
Like this:
Ironically, these “AI detectors” are themselves AI. So if you think AI is powered by environment-despoiling puppy-grinders then you’re turning the crank every time you click that button, and if you don’t trust AI then you’re trusting AI to tell you whether to trust something.
Yeah. There are various limitations and problems with this feature so I’m just trying it out in a very small way at the moment.
Hopefully by defending the threadiverse against LLM slop content the net amount of ground-up puppies will be lower than if we didn’t.
When I coded this originally, a month or so ago, there were a rash of accounts that were only doing LLM-generated posts but after they were banned a few times they haven’t come back. With luck we won’t need this feature anytime soon and I can disable it.
It designates a response as AI for people viewing.
I suggest to rephrase it as not a question but the action that it performs, like “Tag content as written by AI” (but I don’t know the details, like does it always set it to ON, or is it a toggle, how easy is it to find the current state, like does that automatically refresh on the page after clicking this option, etc.), since it apparently seems confusing to the intended users the way it is currently.
I just used you as a guinea pig. Is what it does apparently.
Oh, my apologies - it truly is a question then! Nevermind, it seems well-named in that case.
I do not know, but I can give a guess: When some certain amount of users have flagged the same post/comment as “probably written by AI”, PieFed warns all of the users viewing the post (/comment) after that.
