169
submitted 6 months ago by ylai@lemmy.ml to c/reddit@lemmy.ml
you are viewing a single comment's thread
view the rest of the comments
[-] DaseinPickle@leminal.space 25 points 6 months ago

How do we keep the bots out of Lemmy? Eventually all the AI rot will spread to the fediverse.

[-] Max_P@lemmy.max-p.me 12 points 6 months ago

Defederate/ban them, defederate instances that don't adequately stop them from their instances.

[-] DaseinPickle@leminal.space 6 points 6 months ago

That might work for now, but AI bots can easily create more content faster than human moderators can go through. We will need mechanisms to prove humanness.

[-] henfredemars@infosec.pub 4 points 6 months ago

Another strategy might be to demand a minimum content quality. Whether trash comes from humans or bots, it’s still trash.

[-] dessalines@lemmy.ml 3 points 6 months ago

They rarely get past the signup form, it's trivial to spot them.

[-] ReversalHatchery@beehaw.org 2 points 6 months ago

That's how you dismantle the current network in it's entirety

[-] agressivelyPassive@feddit.de 8 points 6 months ago

Realistically, you don't. You can't really.

The only way would be to verify each account somehow.

[-] DaseinPickle@leminal.space 4 points 6 months ago

Yea, it will be a hard problem to solve. There’s no good solutions right now. Hopefully someone with the creativity and skill will come up with something.

[-] 50MYT@aussie.zone 4 points 6 months ago

It becomes a pain vs gain problem. How hard do you make it, also balancing the inconvenience.

You could easily force users to enter a one time code via email every 3 months (or more or less time). This would be hard to automate and if you changed it up even more so.

[-] DaseinPickle@leminal.space 2 points 6 months ago

Or maybe enforce login with a hardware token like Yubikey or Nitrokey. I don’t think you could automate that(?)

[-] 50MYT@aussie.zone 2 points 6 months ago

Yeah but how do you get it to someone.

Pain vs gain.

[-] DaseinPickle@leminal.space 1 points 6 months ago

I’m not sure I understand. User would have to buy their own keys and set it up.

[-] henfredemars@infosec.pub 7 points 6 months ago

Are AI bots really the true problem, or is the problem product promotion? Humans can do that too.

[-] DaseinPickle@leminal.space 19 points 6 months ago

Yea, but AI bots do it at an unprecedented scale. The price of generating bullshit is practically 0. There is already cheap AI content every where and people are already getting tired of it. It’s true that human actors has done something similar for a long time, but now AI will make the process so easy that AI junk will be everywhere. Look at Google search results, it’s already filled with blogspam generated by AI. Blogspam was a problem before, but now it’s insufferable. Cheap midjourney “art” is showing up in search results even when you search for real classic artist, you get these ugly AI copies. Generative AI is the ultimate bullshit and spam machine.

[-] henfredemars@infosec.pub 4 points 6 months ago

Thank you for the insightful response!

[-] Tregetour@lemdro.id 1 points 6 months ago* (last edited 6 months ago)

Bots aren't the problem. You mistake the tool for its operator by cursing a bot. If I had a specific purpose for a bot, I'd use one. The difference between my using a bot and a clickfarmer using a bot is that I wouldn't treat Lemmy instances as a resource to be exploited (and ultimately ruined). Such an attitude would inform the how and the what behind my bot use.

The problem as always comes back to bad incentives, bad design and bad behaviour.

[-] Snapz@lemmy.world 6 points 6 months ago

I identify with your problem as we are similarly, human people. Have you thought as I have in the past, "Why not let AI solve this problem for me?" I have used BotRemoverAI.co for this issue and the results have been highly acceptable!

[-] fantasty@programming.dev 5 points 6 months ago* (last edited 6 months ago)

Thank you for your suggestion kind stranger take my gold sir edit this blew up thank you for the award

[-] possiblylinux127@lemmy.zip 3 points 6 months ago

We are a community. The community will find a way as we can do anything. We have faith of the heart and everything.

[-] DaseinPickle@leminal.space 8 points 6 months ago

That sounds more like a prayer than a plan.

[-] Tregetour@lemdro.id 1 points 6 months ago

I can see the measures we currently take for ads (large user-maintained subscribable blocklists) being part of the solution. That, and some form of privacy-respecting POW scheme.

this post was submitted on 25 Apr 2024
169 points (98.8% liked)

Reddit

13435 readers
1 users here now

founded 5 years ago
MODERATORS