Anti-AI fedi pact?

a month ago by damon to c/fediverse

Not looking to debate or stir up anything. But, I will say despite anyone’s thoughts the anti-meta fedi pact at least allowed people to have some information on where instances and people stood on federation with Threads. Some decisions were able to be made based on that information. I think if it doesn’t exist there should be one also regarding AI, which is a polarizing topic especially on fedi. I believe similar to the Threads issue it would allow people to make somewhat informed decisions one way or the other. Are there current lists, sites dedicated to this? If not what are you guys thoughts?

iByteABit 8 points a month ago

The Threads issue was not polarizing at all, almost everyone here was clearly on the side that Meta can fuck right off

AI on the other hand does have serious arguments on both sides. What would the resulting decisions even be?

save

path: 0 23915573, hotness: undefined, score: 8, children: 6

p03locke 5 points a month ago

You're going to have a pretty tough time with DB0. I don't think any of us want AI bots posting fake posts or lazy slop bullshit, but many of us just see LLMs as a tool. It's not the technology's fault that a bunch of rich assholes want to drive the economy to the ground over it.

save

path: 0 23915573 23919237, hotness: undefined, score: 5, children: 4

iByteABit 4 points a month ago

Yeah the criticism tends to go to extremes here sometimes. I remember one open source project being criticized just because some AI assisted commits were added and people were saying that it should be forked and shit just because of that.

I'm definitely very against the way companies try to shove it down our throats and make use of it in ridiculous ways and destroy the environment by putting no limits to its use, but the problem as usual is the socioeconomic system behind the tool and not the tool itself.

save

path: 0 23915573 23919237 23920079, hotness: undefined, score: 4, children: 0

CelticViking 3 points 17 days ago

Yep, and local server LLMs that are open source and not corporation-controlled are awesome to use. But Lemmy isn't good a nuance so a lot of posters don't realize that and just scream "ai slop!" at everything. lol

save

path: 0 23915573 23919237 24123818, hotness: undefined, score: 3, children: 2

p03locke 3 points 17 days ago

No mob is good at nuance, which is frustrating. Black and white mentality bullshit.

save

path: 0 23915573 23919237 24123818 24123988, hotness: undefined, score: 3, children: 0

julian 1 point 17 days ago

There's the problematic fact that open weight LLMs are still trained by these corpos, which definitely has some environmental impact. It's troubling.

However what doesn't? As long as I'm not actively burning absurd amounts of fossil fuels running queries...

save

path: 0 23915573 23919237 24123818 24124285, hotness: undefined, score: 1, children: 0

patatas 3 points a month ago

It's totally possible that in your corner of the fediverse it wasn't controversial to defederate Meta/Threads, but it definitely wasn't a consensus position overall.

For example, the largest Mastodon instance didn't defederate, and a lot of prominent voices on the rest of the microblogging fedi argued against defederating. I was pretty shocked about it tbh!

save

path: 0 23915573 23921134, hotness: undefined, score: 3, children: 0

julian 4 points a month ago

It's tough to know where to stand when it comes to AI. I don't think anybody here is hoping that the fediverse be overrun with LLMs and bring about the dead internet theory.

A large part of why people join the fediverse (and the threadiverse, too) is to get away from corporatization, and AI is definitely part of that too.

I just don't think we need to go to extremes to ban AI-generated content, but more importantly, this is not a decision that can be made globally. It must be done (and enforced, however we do that) instance by instance.

save

path: 0 23922295, hotness: undefined, score: 4, children: 0

Zerush 3 points a month ago

I think that it is usefull to ban pure AI generated content created as main information, but allow as contrasted additional information or summary when correct with sources, which can be usefull in scientific documents or long texts. As said before, AI is here and is to us to use it (or not) as tool and with common sense.

save

path: 0 23922656, hotness: undefined, score: 3, children: 0

azdle 3 points a month ago

Not federating with meta was a choice that could be made by the admins, just changing a setting in the instance. Slop can be posted to any instance by anyone with an account. If an admin decided to sign an anti-AI pact there's no way they could realistically abide / enforce it.

save

path: 0 23917421, hotness: undefined, score: 3, children: 2

patatas 3 points a month ago

Not trying to use AI as a moderation tool is definitely up to the admins & mods.

As for a content policy, it could work just the same as any other. Instances have rules against all sorts of content whether it's bigotry, espousal of certain ideologies, abuse, and so on. User posts these things, they get banned by the instance. I don't see how this is different in a technical sense.

save

path: 0 23917421 23921052, hotness: undefined, score: 3, children: 0

libre_warrior 3 points a month ago

They could make a rule saying:

First hand posting of slop is not allowed. Second hand posting is allowed as long as it is sufficiently relevant.

save

path: 0 23917421 23917865, hotness: undefined, score: 3, children: 0