Twitter Can Only Lose When It Polices Abuse
Less than a year ago, in those innocent, dewy-eyed days before Donald Trump had become the front-runner for the Republican nomination, I wrote about an issue that seemed important at the time: Twitter’s harassment problem. Specifically, I wrote about how difficult it was going to be for Twitter to get a handle on its trolls without alienating portions of the user base.
“Alienating portions of the user base” is a bad idea for social media companies that thrive by expanding their networks, not by eliminating users. That's why abuse puts Twitter in a bind: it can lose users by banning racist and sexist accounts, or it can lose users who don't think enough racist and sexist accounts are being banned.
However, the issue has taken on a certain political valence. The users demanding redress seem mostly to be left-leaning people who framed this as a matter of race and gender discrimination. Acceding to their wishes was likely to alienate conservatives, who were likely to experience a disproportionate share of the banning.
After all, what is abuse? There are some things we can all agree are abusive. And then there’s heated rhetoric, sarcasm and all manner of other rude behavior. You can’t police all of it. But policing just part of it -- the part that maps onto identity politics -- comes with its own problems.
It’s easy enough to forbid landlords, employers, and merchants to discriminate against people on the basis of race, gender, sexual orientation, and so forth. But when it comes to speech, those hard-and-fast categories get decidedly fuzzy, because of course, these identities are themselves the focus of political movements. And to state the obvious, the people most active in those movements tend to be folks who identify with the category.
There’s no easy metric for deciding when impugning the characters, motives and intelligence of, say, feminists, crosses the line from a decidedly common political tactic into abuse of women. If Twitter or another network sets itself up to be especially vigilant for users who are at odds with these movements, there's a real risk of declaring open season on conservative ideas, while liberal ideas about a lot of things are carefully protected by your abuse standards.
And indeed, that’s exactly what conservatives think is happening now. Robert Stacy McCain has had his account banned, just weeks after Breitbart’s Milo Yiannopoulos lost his “verified user” check mark. Both practice a brand of Twitter activity one might politely call “obstreperous.” No one knows precisely what led to the disciplinary actions, but McCain believes it’s because of his vocal comments about social justice topics.
We do not, in fact, know that angry, loud conservatives are more likely to be disciplined by Twitter than angry loud lefties; maybe there are a bunch of social justice warriors quietly getting their accounts taken away at the same time for similarly nasty comments. We don’t know whether Twitter’s actions were unjust.
The point is that it doesn’t matter. The moment that you set yourself up as the arbiter of what constitutes abuse, it is virtually guaranteed that you are going to have to make gray-area judgment calls that will alienate some portion of your user base.
When your user base has stalled despite your best efforts to grow it, and investors are getting restive, it’s probably not the right time to signal to half of the political spectrum that their views are not necessarily welcome on your platform. But Twitter didn't have much choice. Aggressive banning is seen as an attack on the right, but continuing the old approach would have been seen as an attack on the left. In the culture wars of today, there’s precious little middle ground.
It is obligatory in accounts of this contretemps to state two things: that the author doesn't like the tactics of the Twitter users in question, and that Twitter is a private company that has a right to ban users. Both true.
For that matter, even strict, unbiased, rules-based bans run that risk. Over the years, for example, I have fielded countless accusations of bias from liberals and social conservative commenters who had fallen afoul of my spam filter. Since they didn’t see the laudatory comments that had also joined them in the never-never bin, they assumed it was personal, and viewpoint based, rather than the accidental result of an algorithm I didn’t even design.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Megan McArdle at email@example.com
To contact the editor responsible for this story:
Philip Gray at firstname.lastname@example.org