Now let us turn to the second category: Twitter censoring speech in order to remove “harmful” content. In 2020, when asked about Twitter’s troubling speech record, then-Twitter CTO Parag Argawal explained that the company’s policy is to focus on the potential harm that a tweet might cause: “We attempt to not adjudicate truth, we focus on potential for harm.”

Perhaps this would be workable if Twitter were to narrowly define harmful speech worthy of restriction in the same way that the Supreme Court has—as “words which by their very utterance inflict injury or tend to incite an immediate breach of the peace.” But Twitter has offered no such satisfying answer to what it actually means by harm. Instead, the labyrinth of Twitter’s rules on the subject is so tangled, broad, and ambiguous that they tell us very little about how Twitter, in practice, determines what speech is harmful.

Since nobody at Twitter seems to know—or at least nobody seems intent on informing the public—exactly how they apply these rules, the company will probably spend years playing language games in order to justify their approach, and eventually come to learn that nobody can agree on the definition of harm. In the meantime, we are left with a standard of harm that is vague, nebulous, and easy to marshall in service of ideological ends. The problem with restricting free speech is not that there aren’t odious opinions; it is that it is naive to trust any institution to determine what they are.