There is and will be an ever increasing demand to filter out AI slop on social media.
Is anyone one nostr working on such a thing?
Discussion
Interestingly enough - I’m out where I don’t have enough reception for my feed to fully load. So I have yet to see any AI Slop, but I’ve read many of your posts about it! 😆
Filters don't work bro, they've said.
I am thinking of other sorts of filtering tho.
You can have a button which is 'mark as ai' at the bottom of posts. You can have some sort of slide on how much you trust other people, or use a web of trust or some such thing.
Perhaps the button is a vote on whether AI or not. Similar to community notes.
That sort of filter.
You'd have to accommodate for abuse of the system of people falsely labeling stuff they don't like as ai or trolling the system. But that could be taken into account somehow.
Could have a report type like the 1984 events but for ai bs. And clients can hide these.
There's even the perfect emoji for marking something as AI slop 🤖
nostr:nprofile1qqsd6ejdteqpvse63ntf7qz6u9yqspp4z7ymt8094urzwm0x2ceaxxgprdmhxue69uhhg6r9vehhyetnwshxummnw3erztnrdakj7qg3waehxw309ahx7um5wghxcctwvshszxmhwden5te0daexc7fdwfjkccte9e5k6ampd3jzuet49u9qq0kn was saying something similar.
I think it's better not to have to do menu diving for this tho, maybe. It should be on the dashboard.
Anyway it's perhaps something a client could experiment with. I think it could be a good USP for one of them.
Yeah ui, and how it's handled In the background are two different things. But yes.
Imagine a Community Vote on whether posts are AI slop or not. That would not only be useful, it would also be pretty fun.
AI will use it to make post more “realistic”
Maybe. But so what.
Besides I don't think there is enough data gathering potential on nostr to be useful lol.
Maybe elsewhere.
Doesn't mean you shouldnt be able to vote on ai slop
I’m just saying. There is more chances I’ll vote in something like this than a presidential election.
Sure I'm just saying that because it could potentially be used to improve realism it doesn't mean it shouldnt be used.
(You could have a slide of confidence of something being AI slop
0-100%)
Well, to me, there is no point. AI slop is being used, a lot, already… even by “reputable” institutions.
Having a label, saying “this was created with AI” has (will have in a near future), the same value as “35mm film” or “oil on canvas”.
Also, ai, is changing too fast, particularly with photo/video. Dalle3 was a huge thing 3 years ago. In 3 years this account could be a fully automated and coherent bot… better than me at engagement. 🪿
I pretty much disagree with every single point in this post.
Also the point of labelling something as AI art is to the have the option for people to mute it
Fair.
Why would people mute something that provides value? Even if it’s ai, or has special effects rendered on blender?
AI slop by it's very definition does not provide value.
I suppose you are talking about wider AI usage. And this brings into question what is slop or not and what is value. But I think you know slop when you see it and I don't think this is a case of one man's trash is another man's treasure.
I can agree with you that my AI semantics is not up to date with “AI slop”
I should just have liked nostr:npub1nxa4tywfz9nqp7z9zp7nr7d4nchhclsf58lcqt5y782rmf2hefjquaa6q8 comment on this thread… and move on with my life.
PS. I will vote on your AI filter 😘
• 79% of the nostr voters has marked this as AI slop
• 5 people you follow marked this as AI slop
i bet you could make a decent ai bot for this. that sounds absurd, but it's actually the opposite of ai slop – instead of expanding low signal into a lot of content, it's giving you a single bit. might not be accurate, but it doesn't waste your time either
No idea how accurate a detection system is, but yeah that could be useful. I'm mostly against the slop, I'm not saying there aren't uses for ai generally.
Hot or not for a new generation. I’m in.
Hi sorry to bother you guys!! We just needed a little boost for our Geyser. it’s for my son 🙏🏻 thanks !! Have a blessed night.
Doesn’t WoT solve that?
Hmm not sure what you mean, web of trust doesn't ensure that people I trust can be trusted not to repost AI slop lol
Mute button? Just mute people with an AI content crutch.
Hopefully it’s only until bubble pops 🦦
That is an option. Some people are AI slop merchants.
I'm thinking more along the lines of it being generally useful, like community notes.
Perhaps AI generated content should be tagged as such by the author or by reviewers/commenters.
That's what I'm saying. It would be like community notes. But I suppose you can't be certain some times so it would have to account for that. And for people abusing the tag.
That’s where WoT would come into play nicely. It would let you see from your followers who did it .
agreed. i thinks it’s weird.
Wait. You are saying that people won't want to filter out AI content? I think you are wrong. I'm not anti AI, but I see no point in using social media filled with ai
I'm saying they will want to filter it out
Oh. That could have been clearer.
I was writing a whole paragraph about how this could kill the internet but as started typing the things I use I could keep using, I realized most thing would be fine. The only things I care it would kill are news and reddit like social media.
I'm not sure the demise of social media would be such a bad thing.
Me neither
Shop has already sailed IMO. Too hard to define and too hard to tell now to know when it is "AI slop". Would be very subjective and not that useful or generally missed.
Totally disagree.
Are you defining "AI slop" as all AI content on social media or a specific subset?
It's a specific subset, but a substantial percentage. I just thought it was not all that interesting to get into this aspect of it, cos it ends up being 'everything is subjective, there is no slop' which is obviously not true.
I was thinking a more broad flag for AI or not (which I think would be useful, but too difficult to reliably implement)
I guess I just struggle to see many examples of AI slop that A) couldn't be marked as spam and B) would be worth spending time to develop a dedicated filter for this specific type of slightly annoying post.
Would I use it if it existed? Maybe. Do I think it's important. Not really. Of course maybe I will change my mind next week haha
This type of ai slop is a huge problem already on most large social media platforms, and it'll only grow. It's already a significant problem. Perhaps you have luckily just not seen it