Follow

Instagram has displayed "false information" warnings with fact checking for a while now.

Is this something you'd like to see on Pixelfed?

help.instagram.com/38853495208

I'm kind of surprised that No is leading.

Thank you for voting and providing feedback, based on the poll results it's unlikely we will pursue this feature.

Show thread

@dansup Yes as long as you open about how "false information" is decided.

@dansup it'd probably be better for false information to just be against TOSes...

@dansup Well, I'm part of a photo/image network to see/post photos and images, not to consume news. :D

@dansup this is an interesting question because my automatic response was yes.
However then comes the question of who or what decides if information is false?
There have been increasing instances of people spreading false information by posing as fact checkers themselves
Who fact checks the fact-checkers?

@dansup It's a moderation feature that should be avaliable

@dansup At @Eunomia our philosophy is that we don't want one entity to be in charge of deciding what's true or not, but rather that users should have tools that simplify detecting misinformation. You should check out our work there. Not saying simply stating something is false is a wrong thing to do but it will put you in a difficult position for sure.

@Gargron @Eunomia A feature like this should be used in extreme circumstances depending on the user/situation in my opinion.

I do understand the risks for abuse though, and federation makes it even more complex.

@dansup
@Gargron @Eunomia

Since I don't know how far you've looked into these topics, I want to ask if I can add the following here:

mastodon.xyz/@d599f84e/1030422

If you have considered the points he talks about, you can ignore my reply here.

@Gargron just asking what is this eunomia thing. it makes me scared.

@Gargron i don't know what this is and it makes me scared

@Gargron because data seems to be involved? scoring systems maybe? i'm not sure what this is and i really want to know the role of eunomia in the mastodon ecosystem

@koyu @Gargron Facts aren't always binary. Sometimes facts can oppose each other and sometimes even major news outlets make a mistake. Even intelligence agencies have made similar systems and failed in doing so.
I doubt you can solve this problem by means of algorithms.
It's up to people what to do with given information and how to process that information themselves.

Also I'm really scared this becomes similar to "reddit up/downvoting" which is based upon personal opinions instead of facts.

@marlon @koyu The first paragraph is exactly @Eunomia's position on the issue and informs our approach.

@koyu Data: You need to be on a Eunomia-enabled server, which does not federate. So this is a strictly opt-in thing.

Scoring systems: There's sentiment analysis and trust voting by users on a per-post basis.

Role in the Mastodon ecosystem: I'm going to say none? It's a standalone project that does not flow into Mastodon development in any way.

@Gargron thank you, that's what i wanted to know. good stuff is safe then :)

@Gargron @koyu I realise it's still early days for #Eunomia, but the description on the site is pretty vague. Looks like "we're going to solve $problem with blockchain!". Making it a system that can't be gamed or abused will be difficult.

The word Eunomia doesn't help either. Obviously Eugen is a fine name, but the eu prefix, meaning good, has some problematic associations.

@Gargron @koyu I realise it's still early days for #Eunomia, but the description on the site is pretty vague. Looks like "we're going to solve $problem with blockchain!". Making it a system that can't be gamed or abused will be difficult.

The word Eunomia doesn't help either. Obviously Eugen is a fine name, but the eu prefix, meaning good, has some problematic associations.

@y6nH @koyu The blockchain in this case is just IPFS and it ensures decentralized storage of the underlying data.

I don't know if rebranding is possible at this stage but I'll pass the feedback on.

@Gargron @y6nH yeah, i mean i slightly cringed at the word "blockchain" when i saw ipfs, but i am not even mad at it, if it says p2p the whole blockchain argument is useless so i'm fine with this...i just wonder how it's decentralized without any servers when there are still firewalls at people's homes in place 🤔

@koyu @y6nH The IPFS will not be used client-side, as I mentioned you need a Eunomia-enabled server, that means all the processing and IPFS storage happens server-side. The decentralization would then be between different Eunomia nodes operated by members of the project like University of West Attica, Greenwich, etc.

@Gargron @y6nH so it's not really a decentralized thing then? more a couple of different authorities running their own thing so no one can alter things i guess? 🤔

@dansup I think such a system comes with great burden to administer and also great power to the admins. I think copying that feature from Instagram would mean we didn't learn from their mistakes.

Maybe we can find a different approach. I like the ToS thought. The bottleneck anyway probably is moderation - and that you need anyway when shitstorms and trolls start hitting the platform.

@dansup no. Fact checking is a centralized approach. Each Armin/mod would need to make decisions.
Algorithmitic approaches (e.g. checks against wikipedia) can easily be patched out by other instances.

@dansup it may be useful in the far future but I don't think it's something you should waste your time on right now. Comments calling people out should be enough for now

@dansup Pixelfed shouldn't try to "mark misinformation" it should moderate properly. And such moderation should include a rule similar to "don't spread harmful misinformation" !

@dansup Could be fine if this will be an opt in settings that can be enabled by the end user.

@dansup I second what @Gargron said. This only has a marginal effect, since someone posting misinformation will probably go to another instance and publish there. Something like a service for reporting fake news in toots and then having an integration inside Pixelfed would be preferable in my opinion. But maybe this is a longer term project.

@dansup
We could put like Wikipedia

"Controversial subject"

**sorry for my bad English**

@dansup

Good question.
Spreading false information should be against the TOS to begin with.
But I can see how users who stumble upon false information could find a link to somewhere explaining why the information is false more useful than just seeing the content disappear/being removed by mods.

@dansup Please do not take my critic personal.Of course i appreciate your personal commitment, but Pixelfed has not reached a mature level of development to think about such a very special functionality. I have said it in the past already, that the development should be much more focused on the basic functionality and a good UX. E.g. why video support ? Wasn't it planed as image application ? As user I do not want to use Pixelfed at the moment, because the usability is not well rounded yet.

@dansup

And who will decide what is false ?
I voted yes because if it's the instance admin it's in accordance with the whole Fediverse concept; but the source of trust has to be decentraized.

I'd like to see any user being able to set up a fact-checking service (e.g. using the tad #factcheck), and any other user to subscribe to whatever fact-checking serves they want to.

I also think that curation/blocklists should happen the same way.

@cabalamat I want a fediverse where all users on it are equal, with the same rights. Not one where we have digital overlords e.g. Twitter, Facebook, YouTube, etc.

@dansup if the media is harmful, the media should be taken down

we're not free speech fetishists around here

@ben it's not that, more of untagged gore warnings, nudity and such. This is an extension of content warnings that are applied by instance admins on a per post basis.

@dansup @ben

That sounds like a button to call someone a liar.

@dansup

@dansup I answered yes, but an ability for instance admins to block posts that are mostly text would do most of the job too.

Facebook/Instagram's problem is mostly that they are loathe to remove infringing posts by the powerful

@dansup I would rather prefer some phone app for pixelfed.social tbh

Sign in to participate in the conversation
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!