There’s nothing that can be done to stop it, but you can downvote or block things to try to reduce their reach.

We cant.

/thread

It’s much worse if you start paying attention to text posts.

Report it so the community moderators can remove it

Not possible

We can do what other spaces don’t, and choose to not design our space around bots in ways that hinder the user experience.

I just go back to the things that forum admins have done forever: block whole IP continents you don’t see meaningful engagement from but see a ton of bots from. Make new accounts jump through a bunch of hoops. Don’t allow new users to create content for a while, and then make them earn that right over time. Shadow ban the crap you can identify so they waste their strength. Reap inactive accounts periodically. And so on.

Get admins with the time and the balls to go after them. Aside from that, nothing we can do as users.

Lemmy won’t like the answer but it’s decentralised anonymous ID (proof of personhood).

Keep normies away from Fediverse. Easiest way. Bad shit follows popularity.

Text wise, there is no way to tell anymore unless the bot fucks up, which happens rarely.

Image wise, I doubt someone has an automatic pipeline where a bot generates pictures on it’s own and posts them. What you are seeing are real people using AI to make their memes, or people who are enthusiastic about AI reposting pictures they find. Nothing wrong with that.

There’s also posting bots, which should be labeled as such but I guess most people don’t. It’s just a way to bring content in, I don’t really see a problem with this.

And there’s vote manipulation obviously.

All that to say that the first and last one are the only problematic ones since it’s used to manipulate and sway opinions. But there isn’t anything we can do if someone is mildly smart about it and uses a proxy service. Llms are simply too good at making themselves look like regular commenters.

We could ask for ID from every user and cross reference with the help of governments to make sure it’s a real one, or have heavy and intense captchas, but who wants to do that?

You can reduce bot noise on the Fediverse through a mix of server settings, moderation tools, and user-side filtering. Since it is decentralized, no one can stop bots everywhere, but individual servers and clients can limit how much spam actually reaches people.

Server admins can require email verification or CAPTCHAs at signup, use manual account approval for new users, limit posting speed for new accounts, and block or silence servers that are known sources of spam. Many Fediverse servers already share blocklists and coordinate moderation so that problem servers get isolated quickly.

Fediverse software is also adding better tools for detecting automated accounts, labeling bots, filtering low-quality AI images, and helping moderators review suspicious posting patterns. Some servers use anti-spam plugins or machine-learning filters to automatically flag or quarantine obvious bot posts. Individual users can mute keywords, block accounts, report spam to their server admins, or switch to a Following-only timeline to avoid noise from the wider network.

Bot spam will never be fully eliminated, but stronger moderation tools, shared blocklists, and user controls make it possible to keep timelines clean without centralizing the network.

Server-side blocklist, auto-synched? Of course always a cat and mouse game.

but now I am starting to see it myself, blatant bots posting really crappy AI images

There’s one account that mods a shit ton.of AI image communities…

If you block that one account, it will.pretty much fix all of it. They’re super weird and going off memory they mod an insane number of communities for AI images, so I wouldn’t worry about pre-emptively blocking everyone, because it’s just the one account posting most of it.

But as you see posts block the community.

It ain’t that difficult.

But there’s entire instances that the admins have lost it to AI brainrot already. That’s where the communities are usually hosted and no amount of facts, logic, or even opinions and feelings will get those admins to “do something” about it.

The fediverse isn’t centralized, you’ll never “get rid of” any subject or topic. Because even if the big instances do, anyone can make a new one.

“Fighting” the people who’s brains have been rotted by AI will never be productive and most likely drives them further into AI

Not sure why it still needs repeated so much, but:

Don’t feed the trolls

Block them and move on.