Where will the company draw the line between acceptable and unacceptable speech?
Twitter appears to have blocked the accounts of several right-wing users, including some associated with white-power groups, as part of a move to respond to user frustration with hate speech and harassment. But it threatens to draw Twitter further into a quagmire over what is acceptable speech.
According to a number of news reports, users who have been banned include Pax Dickinson—a former Business Insider executive who was let go for making racist comments—as well as Richard Spencer, who runs a so-called “alt-right” organization called the National Policy Institute, which advocates for racial separation.
As a matter of policy, Twitter twtr doesn’t comment on actions taken with respects to specific accounts. But the company has taken a number of steps recently to try and rein in bullying and harassment on the service, including rolling out an expansion of its “mute” feature.
Twitter’s co-founder and part-time CEO Jack Dorsey said recently that abuse “has no place on Twitter” and he intends to stamp it out.
There has been a steady drumbeat of criticism aimed at Twitter and its failure to take action to stop hate speech on the platform, apart from certain special cases, including an incident in July when black actor and comedian Leslie Jones was targeted by racists.
After Jones said she was quitting the service due to the unrelenting harassment, Dorsey reached out to her. Shortly afterwards, Milo Yiannopolous, the technology editor of right-wing site Breitbart News and a proponent of various “alt-right” views—including the need for races to live separately—had his account permanently banned.
Get Data Sheet, Fortune’s technology newsletter.
The problem with this approach, however, is that it can have unpleasant side effects. For example, Yiannopolous has said that being banned was the best thing that ever happened to him. Why? Because it fueled his reputation as a critic of the mainstream, and also because it gave him ammunition to argue that he was being targeted by Twitter for telling the truth.
In a similar way, Spencer is already using Twitter’s ban as a sign that he has struck a nerve, and as evidence that the service is caving in to left-wing groups and political correctness.
“This is corporate Stalinism,” he said in an interview with a right-wing site. “There is a great purge going on, and they are purging people based on their views.”
From a historical point of view, Twitter’s problem is that it has always stood for freedom of speech. And over the years, it has gone to considerable lengths to protect the rights of its users to say pretty much whatever they wish, including fighting a French court case that was designed to identify users who posted anti-Semitic and homophobic remarks.
Donald Trump outsmarted Hillary when it comes to earned media. Watch:
But just as Facebook is having to confront criticism about the network’s complicity in spreading fake news—news that may have affected the election of Donald Trump as president—Twitter is having to backtrack from its commitment to unrestricted free speech.
While Facebook has so far consistently refused to implement tools that block fake news (and some, including technology analyst Ben Thompson of Stratechery, argue that it should not), Twitter is now stuck between a rock and a hard place, trying to determine what is permissible speech and what is not.
Perhaps outright racist remarks like those made towards Leslie Jones seem like an obvious candidate for removal. And maybe people can agree that accounts belonging to white-power groups shouldn’t be allowed to remain. But where is the line between a political group that advocates outright racism and the kind of remarks Donald Trump has made about Mexicans and Muslims? Or Breitbart News?
Critics have argued that Twitter only allows such behavior because it is desperate for engagement and user growth, which is similar to the argument for why Facebook doesn’t care about fake news. There is probably some truth to that, given Twitter’s financial status.
But if the company is going to start removing accounts belonging to anyone who says anything remotely offensive, it is going to be spending all of its time doing that, and by doing so it is probably going to alienate as many users as to which it appeals. Do we really want Twitter to be the one that decides what constitutes appropriate speech, and who is allowed to exercise it?