site banner

ACX: Moderation is Different from Censorship

astralcodexten.substack.com

A brief argument that “moderation” is distinct from censorship mainly when it’s optional.

I read this as a corollary to Scott’s Archipelago and Atomic Communitarianism. It certainly raises similar issues—especially the existence of exit rights. Currently, even heavily free-speech platforms maintain the option of deleting content. This can be legal or practical. But doing so is incompatible with an “exit” right to opt back in to the deleted material.

Scott also suggests that if moderation becomes “too cheap to meter,” it’s likely to prevent the conflation with censorship. I’m not sure I see it. Assuming he means something like free, accurate AI tagging/filtering, how does that remove the incentive to call [objectionable thing X] worthy of proper censorship? I suppose it reduces the excuse of “X might offend people,” requiring more legible harms.

As a side note, I’m curious if anyone else browses the moderation log periodically. Perhaps I’m engaging with outrage fuel. But it also seems like an example of unchecking (some of) the moderation filters to keep calibrated.

15
Jump in the discussion.

No email address required.

First, on a personal note, this is exactly what I stoner-hot-take predicted Musk would do with Twitter in a prior motte thread. This freaks me out. Not that it's all that creative a take, but it's something I've noticed before when I was spending too much time in narrow epistemic corners (team fan blogs, fashion blogs) where I'd start to think the same thoughts that showed up on the blogs a week later. The same trade ideas for fan blogs, or I'd pick something up at a thrift store that caught my eye and a week later it would get anointed a trend. It's sort of a weird hivemind thing, we're all thinking about the same issues based on the same influences at the same time if we're all consuming the same set of blogs and news sites. So you can look there for my joking-not-joking predictions as to how this will go, it's a good plan but it won't survive contact with the enemy (users).

Second, to address a specific point SA makes:

A minimum viable product for moderation without censorship is for a platform to do exactly the same thing they’re doing now - remove all the same posts, ban all the same accounts - but have an opt-in setting, “see banned posts”.

Twitter would be completely unusable without any bans or filters, filled with bots and scams and obvious harassment. That's not a viable product. The minimum viable product has to filter out enough to make the product usable. I don't want to see "banned" posts, I want to see posts banned for political incorrectness. Maybe. Shout out to the mods of themotte, would themotte be usable in your judgment without that kind of basic filtering?

Any Chinese person could get accurate information on Xinjiang, Tiananmen Square, the Shanghai lockdowns, or the top fifty criticisms of Xi Jinping - just by clicking a button on their Weibo profile.

"Accurate" is a point of contention here. It's not unusual to have certain topics be overwhelmingly dominated by a particularly numerous or energetic viewpoint on the topic, even just something simple like Toronto sports can get weird with sportswriters admitting they softpedaled coverage of the Raptors and Blue Jays because the articles got a ton of clicks from Canadian fans, along with a ton of comments yelling at the writer if they insulted Toronto's honor. If you're any country other than the USA, you risk being flooded by American content and American viewpoints. While I'm not defending Chinese censorship per se, I do think that saying opening up to all sources of information increases accuracy can be disputed.

First, on a personal note, this is exactly what I stoner-hot-take predicted Musk would do with Twitter in a prior motte thread. This freaks me out. Not that it's all that creative a take, but it's something I've noticed before when I was spending too much time in narrow epistemic corners (team fan blogs, fashion blogs) where I'd start to think the same thoughts that showed up on the blogs a week later. The same trade ideas for fan blogs, or I'd pick something up at a thrift store that caught my eye and a week later it would get anointed a trend. It's sort of a weird hivemind thing, we're all thinking about the same issues based on the same influences at the same time if we're all consuming the same set of blogs and news sites. So you can look there for my joking-not-joking predictions as to how this will go, it's a good plan but it won't survive contact with the enemy (users).

You are a real influencer man! Not the job, the vocation. What used to be called a memelord, or a trend setter. Yeah it is more likely that by reading and watching certain things you are primed to think more on that topic and you just did it faster than others, but rule number one of being an influencer is thinking you're the best and everyone listens to you regardless of what reality says.

Twitter would be completely unusable without any bans or filters, filled with bots and scams and obvious harassment. That's not a viable product. The minimum viable product has to filter out enough to make the product usable. I don't want to see "banned" posts, I want to see posts banned for political incorrectness. Maybe. Shout out to the mods of themotte, would themotte be usable in your judgment without that kind of basic filtering?

Without any bans or filters whatsoever maybe, but why would it be unusable with hidden but accessible shit like that? It seems unlikely any social media company would make a filter to differentiate between banned for spam and banned for politically incorrect - even if the supreme court said they have to display them (but can hide them) to comply with the first amendment, the reason they censored the political speech they don't like is to stop others from seeing it, so mixing it in with spam would be the next best thing.

Without any bans or filters whatsoever maybe, but why would it be unusable with hidden but accessible shit like that?

I'm not sure I phrased that right. I'm saying that the proposed unfiltered product isn't viable if it doesn't distinguish between actual pure trash and political incorrectness. So defiltering makes the product unusable instantly, no one would really use it. The filtered product would still be usable, and dominant. There would be a certain value in the existence of a short-form 4chan aspect to Twitter, but that reminds me of another thing I was trying to get my head around:

To what extent are social media platforms interested in excluding groups that habitually organize to break their internal systems from taking part in their social media ecosystem at all to avoid giving them space to organize? If unfiltered Twitter existed, and rdrama-type trolls were allowed to hang out there provided they all had their filters off, one of the things they're going to do is organize sallies out into the filtered world, and they'll always figure out how to grief users in the overworld. Inviting those people to use a dark corner of your platform is inviting the bikers into the bar, even if they just get a table over there they're eventually going to cause trouble.

did it faster than others

Selection bias. I can't even know of times when I hypothetically would have thought of it slower, if I read an article on RiverAveBlues two days before I would have thought "The Yankees should trade for Trevor Story" then I'm never going to think of trading for Trevor Story on my own. It's more like @daseindustriesltd 's idea of modernity as a distributed conspiracy, if we're all educated in the same universities and reading the same blogs and listening to the same podcasts, the same ideas will occur to everyone.

Ah, you read it as minimum viable, I read it as minimum viable. I have talked with many people who would consider Scott's proposal not censorship, so I could see it being instituted. And I think social media companies would comply with that and stop there. Then everyone would talk about how they love free speech on the filtered version and just treat you like a criminal when they hear you use the unfiltered version. You would be told the information is free, and it's up to you to go find it, which is true.

Selection bias. I can't even know of times when I hypothetically would have thought of it slower, if I read an article on RiverAveBlues two days before I would have thought "The Yankees should trade for Trevor Story" then I'm never going to think of trading for Trevor Story on my own. It's more like @daseindustriesltd 's idea of modernity as a distributed conspiracy, if we're all educated in the same universities and reading the same blogs and listening to the same podcasts, the same ideas will occur to everyone.

It sounds like you are describing an egregore? Or maybe the zeitgeist as an egregore? That's one way you can view history, as a bunch of different egregores fighting it out on the conceptual plane. First there were family egregores, then tribe then village and so on (although all those smaller ones remained). You are noticing your integration into those egregores I think, and as a trend setter that's how it looks. As a regular member it just looks like everyone is saying on fleek all of a sudden for no particular reason.

Thanks for turning me on to tht term egregore. That's an interesting rabbit hole.

They are quite fascinating yeah. Make sure you search the vault here too, I think some of our smarties have talked about it (although it is a rather old concept and while it has always had mystical elements it has always been sociological or anthropological - the mystical stuff was because we didn't understand memes and so didn't have a scientific framework to use I think.)