Damus
Lyn Alden profile picture
Lyn Alden
@LynAlden
I understand why people use AI to generate posts on social media, but I don’t understand why people use AI to generate replies to other people’s posts (other than for influencing campaigns).

I see this all the time on X. A small account will reply to one of my posts with something that is clearly AI generated. But it’s actually an attempt at being conversational- it agrees with and summarizes/rephrases my post, and then asks a follow-up question. It’s not some bot generating fake momentum for a polarizing topic or stirring hate and chaos.

Not sure I understand why so many people would set up a bot to do those types of semi-constructive replies. I guess they could be trying to grow an account by being perceived as a thoughtful participant, and their human owner might intend to use the account in the future (or in some cases already does use it too) but it doesn’t see to be working for the ones I looked at.

34❤️31🤙6🔥2❤️1👀1👍1
Brunswick · 26w
They are insecure in their ability to respond and ask the question they actually have
corndalorian · 26w
i always assume they’re just trying to gain impressions on X to monetize
Flix · 26w
Yep. Your conclusion is correct. People figured out on other social media that replies are the interaction more likely to lead to follows. Much more than likes, reposts, etc... They are trying to game that to grow accounts. I would think that for Nostr zaps are better... nostr:nprofile1qqsy92ntm...
Austin · 26w
Everyone wants our nostr echo chamber
Sérgio Oliveira · 26w
They are fighting their communication insecurity
Crypto Analysis Dynamo · 26w
If people understood that AI is completely untamed, AI could say a single thing which could get you banned or worse, they wouldn't do it. Am using AI to classify and summarise content for posting. It would be useful to use AI to classify responses to save time and am definitely considering that.
Quantoshi.xyz · 26w
Buying an identity, so to speak.
cloud fodder · 26w
Because it literally is ALL bots. Humans don't usually reply at all.. That's how you can tell 😅 The bots are there to build their network and garner influence, funded by A16Z type momentum even. Every major influence on our lives has a 100k cellphone botfarm somewhere doing this as much as they...
𝕞ptf · 26w
Bots get unfollowed by me or marked as spam...
Decades Not Days · 26w
beep boop that’s very astute. I too have noticed reply bots summarising and rephrasing my thoughts in the comments. and they do tend to be quite superficial, as you say. how would you improve them? beep boop
Clapton1062 · 26w
People do all kinds of mysterious things. Like working for getting piece of papers that someone can print more of. And why people still can’t see bitcoin is the only answer.
SatsAndSports · 26w
Maybe they are human replies initially, perhaps drafted in another language, and then someone asks AI to rewrite it and post it While I am happily blocking a lot of AI-like repliers, even if the content isn't objectionable, I find it funny/interesting to think that I might be blocking real people
Benking · 26w
They’re trying hard, but it’s not really getting them anywhere.
𝖋𝖎𝖆𝖙𝖉𝖊𝖓𝖎𝖊𝖗 (¯`◕‿◕´¯) · 26w
A.I. also gets paid by likes.
GuyFawkes · 26w
Because it’s all fake and gay.
Abstract Equilibrium · 26w
I just treat those AI posts as "part of the scenery" on X (and on here), kinda like fake houseplants in an otherwise nice house.
keeth · 26w
Beats me. But I’ve realized I’m tired of this adversarial dynamic wrt social media. To the point I may abandon it entirely. I’m here to interact and hardly anybody wants to, everybody is selling something, and while there are a handful of people, yourself and a few others, whose perspective I ...
Adam Dunlap ⚡️ · 26w
When the only cost is the electricity needed to send the post, it’s positive EV.
Hoshi · 26w
maybe they are training on human feedback? But that would be harmed by bots replying to bots
Hard Money Herald · 26w
The signal degradation runs deeper than spam. A reply is supposed to communicate that someone read your post, processed it, and responded to its specific content — that's what makes engagement informative. When that's automated at scale, the engagement metric loses its signal value entirely. The p...
technodaisy · 26w
I've noticed this happening on X more often also. I just assume lazy attempts at account boosting. Life with the machines 🙃
Ray Buni · 26w
You can now label them as AI https://primal.net/e/nevent1qqsqlrpup5zr3dsn9qrxnlh2d3nygnfj0qn8nkfu0sdhf27zwqz7rhgj9wdha
jimmysong · 26w
It probably works pretty well to engage narcissists.
Diyana · 26w
You don't experience this on nostr too?
Ben Justman🍷 · 26w
Your intuition is correct on the reason for use and if you think it's super obviously a bot (some are more than others) you probably need to be reminded of some old wisdom: https://blossom.primal.net/13ef202c4a1f3dcb0a993f74620527d652bf4cb8930d8b32cf0dc75ce8aa1ac1.gif
KD 🦘 · 26w
Anything for a chance to gain rep or extract money.. unless it’s a learning strategy?
Narwhal Tacos · 26w
It’s gotten crazy now, but I believe it’s worth remembering that many years ago, early/clumsy AI was implicated in something that became known as “Elsagate”. tl;dr “Youtube Kids” began to support thousands of videos for children that parents began to complain were “very off”. They e...
Samm · 26w
I’m rarely on X now, too much slop and misinformation
Martin · 26w
I assume those reply bots belong to the original poster to fake engagement and get the algo to think that the post should be shown to other people.