this post was submitted on 04 May 2026
340 points (90.7% liked)

Fediverse

41908 readers
366 users here now

A community to talk about the Fediverse and all it's related services using ActivityPub (Mastodon, Lemmy, Mbin, etc).

If you wanted to get help with moderating your own community then head over to !moderators@lemmy.world!

Rules

Learn more at these websites: Join The Fediverse Wiki, Fediverse.info, Wikipedia Page, The Federation Info (Stats), FediDB (Stats), Sub Rehab (Reddit Migration)

founded 2 years ago
MODERATORS
 

I recently discovered that some popular federated instances have been using LLM-assisted moderation tooling that evaluates whether someone has said something bannable. They do this by running a script/app that sends the user’s comment history to OpenAI with the question “analyze this content for evidence of specific political ideology sentiment. Also identify any related political ideology tropes“. (The italic bits are where I've redacted the ideology they're seeking).

OpenAI’s LLM (they’re using GPT-5.3-mini) then responds with something like:

image

and so on, hundreds of comments.

I have not named the instances or people involved, to give them time to consider the results of this discussion, make any corrective changes they want and disclose their practices at their own pace and in their own way. I have also redacted the evidence to avoid personal attacks and dogpiling. Let’s focus on the system, not the individuals involved. Today these instances and people are using it and maybe we’re ok with that because it’s being used by groups we agree with but what if people we strongly disagree with used it on their instances tomorrow?

The use and existence of this tooling raises a lot of other questions too.

What are the risks? Fedi moderators are often unsupervised, untrained volunteers and these are powerful tools.

What safeguards do we need?

Would asking a LLM “please evaluate this person’s political opinions” give different results than “find evidence we can use to ban them” (as used in the cases I’ve seen)?

What are our transparency expectations?

Is this acceptable and normal?

Should this tooling be disclosed? (it was not – should it have been?)

If you were given a choice, would you have opted out of it?

Can we opt out?

Are there GDPR implications? Privacy implications? Should these tools be described in a privacy policy?

Are private messages being scanned and sent to OpenAI?

How long should these assessments be retained and can we request to see it, or ask for it to be deleted?

Once the user’s comments are sent to OpenAI, is it used to train their models?

What will the effect be on our discourse and culture if people know they are being politically profiled?

Where are the lines between normal moderation assistance tools, political profiling and opaque 3rd-party data processing?

I hope that by chewing over these questions we can begin to establish some norms and expectations around this technology. The fediverse doesn’t have any centralized enforcement so we need discussions like this to develop an awareness of what people want in terms of disclosure, privacy, consent and acceptable use. Then people can make choices about which instances they join and which ones they interact with remotely.

And of course there are the other issues with LLMs relating to environmental sustainability, erosion of worker’s rights, increasing the cost of living and on and on. I can’t see PieFed adding any functionality like this anytime soon. But it’s happening out there anyway so now we need to talk about it.

What do you make of this?

top 50 comments
sorted by: hot top controversial new old
[–] forestbeasts@pawb.social 3 points 1 hour ago

Oh fucking YIKES.

Do NOT send our post history straight to OpenAI, that's just ... extremely gross.

Sure, it's "public", but that doesn't mean feeding it directly to the slop machine is okay.

-- Frost

[–] Auth@lemmy.world 1 points 6 hours ago

The instances doing this were awful before this and will continue to be so with or with it. Its not something i would ever want to see from my instance.

[–] General_Effort@lemmy.world 4 points 11 hours ago

GDPR-wise, this is the absolute nightmare scenario.

Data about the political orientation is defined as especially sensitive ("special category data"). When people just straight post their ideological leanings, that's one thing. But what's described here is profiling. All the available data relating to a person is analyzed by "automatic means" and used to assess their leanings. This then is used to discriminate against them. It doesn't get much worse.

This might be legal in very specific circumstances. EG non-profit religious or political organizations are allowed to police their members and associates to some degree. That would involve quite some extra paperwork. But it doesn't apply here anyway.

Apparently that is on top of ordinary GDPR violations. The processing is done by a third party (OpenAI) without the necessary paperwork. You remember that billion Euro fine that Meta got? That was because they processed data outside the EU, in the US. And that wasn't even "special" data.

You know how those cookie banners in the EU look like? That's for normal data. All the disclosure, all those settings are legally required. Some people on the Fediverse go apeshit over far smaller things.

This may also be a problem for other instances. Your instance sends all your data (except e-mail and IP address) to anyone in the world who asks, with no strings attached. That may be okay as long as users understand that that's exactly what they sign up for. Looking at comments here, it doesn't seem like that is universally understood. That's a problem. On top of that, we now have a situation where there are hints that the personal data is being abused.

[–] skisnow@lemmy.ca 14 points 18 hours ago (1 children)

LinkedIn's LLM-powered automation banned my account on a false positive a few months ago, and it took ages to get it sorted out and they treated me like shit the entire way through even after acknowledging that they'd made a mistake. Sadly it's extremely difficult to operate in my field without a LinkedIn account, because I would love to be able to delete it.

This shit is poison

[–] JackbyDev@programming.dev 5 points 17 hours ago

As trash as LinkedIn is, I've gotten three of my last four jobs through there and a ton of leads and offers. Yeah, I really couldn't imagine getting banned. That freaks me out.

[–] Duke_Nukem_1990@feddit.org 10 points 17 hours ago* (last edited 17 hours ago) (3 children)

Lol it's dbzer0 isn't it.

But also, I don't really care. The fediverse is open by design, you don't even need an account to access the data. I don't like it but we can't really do anything against it.

[–] 7101334@lemmy.world 1 points 1 hour ago* (last edited 1 hour ago)

I care because it impacts the environment to a substantial and unnecessary extent. Any instance using AI moderation should be defederated for that reason.

But personally idgaf about people seeing my data, I wouldn't put it on the internet otherwise. Yeah, morally-ethically that shouldn't be the world we live in, but you have to operate in your paradigm while endeavoring to change it. Still fine with whacking an instance for erring on the side of stupidity though.

[–] patatas@sh.itjust.works 1 points 8 hours ago (1 children)

Even putting aside the distinction between "possible" and "ethical" when it comes to data harvesting and use - just because something is technically possible does not mean it is therefore OK to do - this is moderators using systems created by entities explicitly aligned with US fascism to do "ideology analysis".

If the point of the fediverse is to avoid centralised/corporate control and influence over our relationships, then this is completely antithetical to that goal.

[–] Duke_Nukem_1990@feddit.org 0 points 6 hours ago (1 children)

How so? I'd say it's working exactly as intended. We can just defed the offending instance. This is decentralization realised.

[–] patatas@sh.itjust.works 1 points 3 hours ago

Forgive me, but that's quite a different statement from your previous comment which said "I don't care" and "we can't really do anything"

[–] danhab99@programming.dev 0 points 7 hours ago (1 children)

inb4 if you don't like being modded by AI start your own instance

Personally I'm shocked that this isn't more prevalent.

Reddit was already hard enough to moderate without AI tools. Now in the year of 2026 with what amounts to entirely volunteer-based "companies" or non-for-profits(atleast) running Lemmy instances for us for free you have to get AI help for moderating.

I've been working on a competitor to the activity pub protocol and I have a ready-made solution called userless and the only reason I've never deployed a demo server for other people to test and interact with is because I have no idea how I would moderate it! That's encouraged me to work on the peer-to-peer version of the protocol so I don't have to moderate it at all but still this isn't easy.

And to address the privacy concerns about who is moderating you.... This is the public internet, your data is shared because you share the data. How can you expect privacy in public.

[–] 7101334@lemmy.world 1 points 1 hour ago* (last edited 1 hour ago)

Reddit was already hard enough to moderate without AI tools.

...It's really not. I moderate a few subreddits (a sin, I know, but I'm here too). It's fine. You just wait for people to report shit and look around a bit yourself, same as 2016, same as 2010. Get a handful of people doing that together, or a bigger handful for bigger subs, set up some basic automod stuff for frequent spam links or slurs or such, and it'll be fine.

[–] sobchak@programming.dev 24 points 22 hours ago (3 children)

I'm mostly just surprised that a mod would pay for tokens to moderate. The Fediverse is radically public by design, so I don't have any expectation of privacy. I'd bet at least someone is gobbling up the entire Fediverse to train AI, since companies are so desperate for new human-generated data.

[–] ZombiFrancis@sh.itjust.works 5 points 14 hours ago (2 children)

Their whole thing is that they're running local models on their own systems, so it is unlikely to be corporate at all.

Now, there is a portion of lemmy instances and piefed that did incorporate, and they seem to have it out for db0.

[–] technocrit@lemmy.dbzer0.com 4 points 13 hours ago* (last edited 13 hours ago)

Now, there is a portion of lemmy instances and piefed that did incorporate, and they seem to have it out for db0.

Corporations hating on anarchists? Wow I'm shocked.

[–] sobchak@programming.dev 4 points 13 hours ago (2 children)

IDK, this post says "GPT-5.3-min" which is not an open-weight model AFAIK. Either way, everything is public on the Fediverse, so IDC very much. Specifically asking for political classification is a little bit weird though.

[–] ZombiFrancis@sh.itjust.works 3 points 13 hours ago

I know at least some models can be rigged to local machines, and I assume chatgpt is one of them. But no disagreement that there's a bit of weirdness here.

[–] mathemachristian@lemmy.blahaj.zone 3 points 13 hours ago* (last edited 13 hours ago) (1 children)

It's asking for zionist posts, because that's what's most apparent about the user. That's what they got banned for, db0 is very clear about their "no zionist" policy so it's rather clear that samskara was gonna get banned from there no matter what.

Also the admins say it was a local model that they pirated or smth idk

[–] 7101334@lemmy.world 2 points 1 hour ago

I have to pick between siding with a Zionist and siding with AI users?

Nightmare scenario. Fine, AI users.

It was a local model.

load more comments (1 replies)
[–] Andy_R@feddit.uk 21 points 1 day ago (1 children)

Is Rimu okay lately? He’s been acting so hostile.

I don't think so, then again he hasn't been okay for a while. Him making poor descisions about piefed and deleting users who are critical of him aren't a new thing. They're par for the course. People have criticized piefed for this for a while, but they've been dismissed as tankies or reactionaries. I admit I used to dismiss them as well. Not anymore though.

load more comments
view more: next ›