Harmful is not the same as offensive

If I insult somebody on a forum, the argument goes, how is the forum responsible for that.

It’s a bad analogy and it’s bad in a way worth pulling apart, because the flaw in it is the same flaw running through most of the arguments you’ll have about this. Whether something is insulting can only be decided by the person you aimed it at. The same goes for whether a comment is negative or stupid or whether a derivative work infringes anything. All of those are matters of opinion and the two parties involved will almost always take different positions on them. That’s not a flaw in the people arguing. It’s the nature of the category.

So when you ask a platform to act on that kind of content, you’re asking it to settle something with no fact of the matter in it. Whatever it decides, someone who honestly holds the other view gets overruled. That’s hard, and it’s why moderating speech stays a permanent argument rather than a problem you can expect anyone to solve.

Now change the example. Suppose I invent a piece of malware that’s just plain text and I post it in a comment. Should the platform take it down or leave it sitting there.

That’s a different question and it has a different shape because now there’s no matter of opinion involved. The thing does damage to whoever encounters it and it does the same damage regardless of how anyone feels about it. You don’t have to decide whether you’re offended. The harm is a property of the object rather than a relationship between two people.

My position is that when something is actively harmful in that sense, the publisher has a duty to protect its visitors by removing it or limiting what it can do. Not because publishers should be arbiters of what’s acceptable, which is the thing everybody is quite reasonably worried about but because those are two different jobs and only one of them requires a judgement about taste.

It’s worth noticing that this already happens on the sites you use and nobody objects much. The forums people hold up as bastions of free discussion have moderators who remove things in extreme cases and mechanisms for users to flag content so a consensus can hide it. That’s a form of censorship by any strict definition, it exists on the sites people praise for not having any and it’s mostly uncontroversial precisely because it’s aimed at the objective end of the spectrum.

Plenty of people take the opposite view and think platforms shouldn’t get involved at all. That’s a coherent position and I’m on the other side of it, though not by a huge margin. There are two sides to most arguments and this is one of the ones where the sides are arguing about different cases without noticing.

The useful move, if you want your own argument to survive more than one exchange, is to keep the two categories apart. Ask whether the harm exists whatever anyone thinks of it. If it does, you’ve given the platform something to act on that doesn’t need it to referee your taste. If it doesn’t, then whatever the platform does is going to be a judgement call and pretending otherwise is how these arguments stay unresolvable for a decade.

You can see how much work the distinction does by watching what happens when it collapses. Once a platform accepts that it should act on things people find upsetting, it has taken on a job with no stopping rule and no way to be right, and every decision it makes afterwards becomes evidence for whichever side lost. Once it accepts that it should act on things which do damage regardless of opinion, it has taken on a job that’s difficult and expensive and at least has a shape.

The first job is the one everybody argues about. The second is the one that actually protects anyone, and it gets very little attention because there’s nothing to argue about in it.