Two Truths: Free Speech and the Rules That Govern Online Platforms

How to read this

Before arguing for our own view, we practice stating the strongest fair version of a fellow citizen's view. This piece takes one contested question — how much authority private platforms should exercise over what people post, and how much pressure from government or the public should shape that authority — and gives each of two positions the same space, the same specificity, and the same charity. Both cases are held by people who value free expression and who worry about real harm; they weigh the risks differently.

The case for maximizing open expression

The argument for maximizing open expression begins with an observation: a small number of very large platforms now function as the public square for most political and cultural conversation in America, and a decision by one company's trust-and-safety team can silence a speaker as effectively as a government order once could. Advocates of this view argue that when so much speech runs through so few chokepoints, the traditional remedy for bad speech — more speech in response — cannot work if the bad speech is never allowed to appear at all.

They point to specific patterns. Moderation is enforced unevenly, often because the volume of content vastly outstrips the humans reviewing it, so automated systems flag satire as harassment and miss coordinated harassment entirely. Rules written to stop misinformation have, at various points, suppressed claims that were merely contested and that later turned out to have merit. Advocates note that the people most confident a given rule is being applied fairly are usually the people whose own views the rule was never going to touch.

There is a fairness argument too. If platforms are going to set rules with the reach of national policy, users have a right to know what those rules are, how they are enforced, and a right to appeal a decision to an actual person rather than an opaque system. Open-expression advocates believe that erring toward leaving content up, and countering bad speech with better speech, produces a healthier public conversation over time than erring toward removal.

Finally, they argue, the instinct to moderate away everything uncomfortable underestimates the public. People are capable of encountering a bad argument and rejecting it; a public square managed to prevent that encounter is not healthier. It is only quieter.

The case for responsible content moderation

The argument for responsible moderation also begins with an observation: an unmoderated platform does not default to open debate; it defaults to whoever is loudest, most coordinated, and least constrained by norms of honesty or civility. Advocates of this view point to specific, documented harms — coordinated harassment campaigns that drive people, disproportionately women and members of minority groups, out of public conversation; influence operations designed to look like ordinary citizens; and health misinformation that measurably changed behavior during a public crisis, in some cases fatally.

They point to specific, targeted tools rather than blanket suppression. Removing coordinated inauthentic accounts is not censorship of an idea; it is enforcement against impersonation, which every earlier medium also policed. Labeling content with added context, rather than removing it, lets a platform flag a likely falsehood without deciding the question for the reader. Slowing the spread of unverified claims during a fast-moving crisis buys time for verification without permanently silencing anyone.

There is a fairness argument here too. A platform with no moderation at all is not neutral; it is a design choice that favors whoever can produce the most content, including bad-faith actors with the resources to do so at scale. Advocates argue that the absence of a rule is itself a choice, and that pretending otherwise lets platforms disclaim responsibility for outcomes their own design produced.

Finally, they argue, a private company retains the right, and often the practical obligation to its own users and advertisers, to decide what appears on a service it built and is legally responsible for. That is a different thing from government censorship, even when the effect on a given speaker feels similar.

Where honest people still differ

• Where the line falls between removing coordinated abuse and removing an unpopular opinion.
• How much transparency and due process a private platform owes users over enforcement decisions.
• Whether government requests to platforms to remove content are appropriate coordination or improper pressure.
• How to weigh the harm of leaving a false claim up against the harm of wrongly removing a true one.

What each side could concede

Open-expression advocates could concede that unmoderated platforms have enabled real, targeted harassment that drives people out of public life, that coordinated inauthentic activity is a legitimate target for removal, and that some misinformation has caused measurable harm.

Moderation advocates could concede that enforcement has been applied unevenly and has sometimes suppressed claims later vindicated, that opaque appeals processes erode trust even among people who support moderation in principle, and that a platform too quick to remove contested speech can damage public trust as much as one that removes nothing at all.

Neither side wants a public square ruled by whoever shouts loudest, or by whoever controls the removal button. That much, at least, is common ground.