The ethics of restricting speech on social media

There are good reasons for tech companies to prohibit hate speech on their platforms. But their status as private organizations is not one of them.

Marchers at a white-supremacy rally encircle counter protestors at the base of a statue of Thomas Jefferson after marching through the University of Virginia campus with torches in Charlottesville, Va., USA on August 11, 2017. (Shay Horse/NurPhoto via Getty Images)

Marchers at a white-supremacy rally encircle counter protestors at the base of a statue of Thomas Jefferson after marching through the University of Virginia campus with torches in Charlottesville, Va., USA on August 11, 2017. (Shay Horse/NurPhoto via Getty Images)

As we mourn the death of Heather Heyer, who was run down by a car driven by an alleged white supremacist in Charlottesville, Va., last weekend, many tech companies are reflecting on the role they may have played in enabling white supremacists and other far-right extremists to organize, promote hate and advocate violence. While a number of companies have taken steps to censor the messages of, and refuse service to, extremists, some argue that more needs to be done. Critics worry about the effects of limiting speech and the harm it could do to public political discourse. Is there a case for banning extremist groups from social media and internet platforms? And what should we think about it?

READ MORE: Charlottesville and the politics of fear

A common approach to thinking about the permissibility of limiting speech focuses on the kind of organization considering the restriction and the opportunities and barriers it faces to doing so. For governments, legal and constitutional protections of speech create a narrow space for restriction. Reasonable limits are sometimes imposed—such as Canada’s prohibitions on hate propaganda in the Criminal Code (specifically, sections 318, 319 and 320)—but the standard of justification for those limits is quite high.

By contrast, when businesses and other private actors consider restrictions, they have more room to act. The flip side of Silicon Valley’s libertarian, content-neutral ethos—which permits all kinds of offensive content—is a libertarian, property-rights-protecting disposition—which, arguably, allows them to impose restrictions on content arbitrarily. A case in point is the justification Cloudflare CEO, Matthew Prince, offered in a memo to employees for his decision to boot the extremist site Daily Stormer from its platforms. “The people behind the Daily Stormer are assholes and I’d had enough,” he wrote. “I woke up this morning in a bad mood… It was a decision I could make because I’m the CEO of a major Internet infrastructure company.”

For some, the fact that Prince and other technology CEOs face fewer barriers to banning extremists provides sufficient reason to ban them. But focusing on the nature of the organization considering limits gets us only so far. The challenge is determining where to draw the line between offensive speech and hate speech, and what reasons might be offered for banning the latter and permitting the former. That, in turn, means that we need to think about why speech matters and confront both the benefits and risks of restrictions.

Consider John Stuart Mill’s classic defence of what he called the liberty of thought and discussion. Mill argued that we ought to tolerate offensive opinions because of the benefits such opinions produce for human and social progress. In some cases, an opinion may be offensive, but true—in which case silencing it would rob us of access to the, admittedly uncomfortable, truth of some matter. In other cases, an offensive opinion may be partly true and, when combined with our partly true popular opinion, produce a greater understanding of the whole truth. Unrestricted speech creates an opportunity to learn and correct errors.

RELATED: Democracy can’t be taken for granted. Charlottesville proves that.

But even in cases where an offensive opinion is false, Mill says that hearing and responding to it can help us better understand the reasons why our current views might be correct and justified. If a strongly held opinion or conviction is not “fully, frequently, and fearlessly discussed, it will be held as a dead dogma, not a living truth.” For example, we might develop a better understanding of exactly why slavery, racism and sexism are wrong—and thereby put ourselves in a stronger position to recognize and dismiss arguments for them—by occasionally grappling with the opinions of jerks who think otherwise.

But notice that if the value of liberty of thought and discussion resides in its contribution to individual and social development, then we should be just as conflicted about restrictions on speech imposed by private individuals and organizations as by restrictions imposed by governments and political institutions. The silencing of an opinion itself is the primary issue, according to Mill, not the status or identity of the censor. “Protection, therefore, against the tyranny of the magistrate is not enough,” he writes. “There needs protection also against the tyranny of the prevailing opinion and feeling; against the tendency of society to impose, by other means than civil penalties, its own ideas and practices as rules of conduct on those who dissent from them.” On what grounds, then, can speech be restricted?

Mill famously argued that liberty—and not just liberty of thought, but of action as well—ought to be restricted only when its exercise harms others. “The sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number,” he says, “is self-protection…[T]he only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.” To restrict speech, in Mill’s view, it is not enough that an opinion is offensive; there must be a reasonable expectation that not restricting the message could lead to tangible harm.

READ MORE: It wasn’t a lone, unusual flare-up. Charlottesville really is America.

The challenge is determining exactly what constitutes harm and, therefore, what might justify limits on speech. I take offence when someone calls me an idiot. But while that happens more often than I’d like, I’m not harmed in any meaningful sense. When someone encourages someone else to kill me (which happens less frequently, but still more often than I’d like), there is a case for censoring that message—because it threatens substantial harm. But there is a wide open field between those two cases. Exposure to relentless racist, sexist or other vile messages—short of encouraging violence—might constitute a kind of harm. Or it might not. Distinguishing between cases that are merely offensive and those that constitute harm requires thoughtful deliberation and judgment. Mill offers a useful principle, but not an algorithm, for sorting through them.

In thinking about restrictions, then, both advocates and critics should pay less attention to what private organizations are permitted to ban, and more on what they may or may not have reason to ban. If there is something to worry about when tech companies think about speech and censorship, it is that they have enormous power to shape discourse, but little corresponding obligation to provide a formal account of how and why they shape that discourse. In banning hate speech, they’re on solid ground. But what’s less clear is whether tech companies are equipped to distinguish between genuinely harmful and merely offensive speech. Tech companies and their CEOs need to think about good, defensible reasons for acting, and not simply on how they feel when they wake up in the morning.

Dan Munro is Visiting Scholar and Director of Policy Projects in the Innovation Policy Lab at the Munk School of Global Affairs at the University of Toronto. Listen to The Ethics Lab on Ottawa Today with Mark Sutcliffe, Thursdays at 11 a.m. EST. @dk_munro


The ethics of restricting speech on social media

  1. Fascism requires censorship.

    Knowledge is power and fascists need to control citizens.

    What is democracy when people cannot speak freely?

    Our Supreme Court has ruled that truth can be no defence for those accused of hatred. That means they have NO DEFENCE. It couldn’t be more fascist.

    Truth can NEVER be hatred. Hatred is conflict and conflict in speech arises from a lie. Truth resolves conflict.

    I’m not responsible for how the truth makes you feel. You have NO RIGHT to call the truth hatred to satisfy your corrupt ego.

    • Our Supreme Court has ruled that truth can be no defence for those accused of hatred.

      You mean the SCC? I call BS. Please cite the case.

        • That is not a source…..that’s some idiots blog.

          The Supreme court never ruled any such thing.

          • I hope Keith can read.

          • A weak argument is supported by the stupid denial of documented proof in an attempt to satisfy an ugly ego

          • Dude…..this is Canada….not the US

            Your ‘supreme court’ has nothing to do with us

          • Gotta love free speech.

            The useful idiots demonstrate their weakness, stupidity and ugliness voluntarily.

        • Your choice to link to a screed, rather than any of the more reputable analyses (or to the decision itself) – not to mention the toe of your comments – tells me I’m probably wasting my time, but here goes…

          So… the Whatcott decision. Have you actually read it? I did when it was first released, along with a lot of the commentary.It was somewhat controversial. I’ve taken the time to look it over again in replying to you. Admittedly, I didn’t read the whole thing this time (it’s pretty long) but read the headnote in full, as well as certain salient sections of the decision itself to refresh my memory.

          [For those not familiar with the case, Mr. Whatcott was originally found to have published ate speech by distributing a number of pamphlets maligning homosexuals. The SCC agreed that two of the four pamphlets did indeed constitute hate speech – partially upholding the tribunal’s decision. Full decision is found here: ]

          First, from the headnote: “The words and phrases in a publication cannot properly be assessed out of context, and the expression must be considered as a whole, to determine the overall impact or effect of the publication.” Speaks precisely to what you are doing here, does it not?

          Having said that,I’ll pull some quotes that give context to your claim. First, you seem to be paraphrasing (badly) this bit from para. 136: “The legislature of Saskatchewan has not provided a defence of truth or any other defence as a basis upon which to avoid being found in contravention of s. 14(1)(b) of the Code…” This is a statement on the Sask. Human Rights Code. It’s not saying, as a blanket claim, “truth is not a defence.”

          Let’s take a look at some of the other things said – from para. 140: “Truthful statements can be interlaced with harmful ones or otherwise presented in a manner that would meet the definition of hate speech.”

          Or para. 141: “To the extent that truthful statements are used in a manner or context that exposes a vulnerable group to hatred, their use risks the same potential harmful effects on the vulnerable groups that false statements can provoke. The vulnerable group is no less worthy of protection because the publisher has succeeded in turning true statements into a hateful message. In not providing for a defence of truth, the legislature has said that even truthful statements may be expressed in language or context that exposes a vulnerable group to hatred.”

          So the question was whether or not the statements he made were likely to incite hate. Claiming elements of what he said was truthful didn’t matter, if those elements were couched within phrasing designed to create hatred toward a group protected by the Charter. So what kinds of things did he say? Among other things:

          [189] The flyers also seek to vilify those of samesex
          orientation by portraying them as child abusers
          or predators. Examples of this in Flyers D and E
          would include: “Our children will pay the price
          in disease, death, abuse . . .”; “Sodomites are 430
          times more likely to acquire Aids & 3 times more
          likely to sexually abuse children!”; and “[o]ur
          acceptance of homosexuality and our toleration
          [sic] of its promotion in our school system will lead
          to the early death and morbidity of many children.”

          Seems pretty clear what his goals were: To vilify a group based on their sexuality.

          Ultimately, Whatcott was found guilty of publishing materials likely to incite hate. Not for telling the truth, as you imply. Which leads to my closing question: Why are you looking to distort the truth?

          • *tone of your comments*

            Really wish there was a way to edit comments on here…

          • I don’t know what you’re complaining about.

            Anyone who takes the time to read your analysis will see that it supports my statement.

            Specifically para 141.

          • It shows that the court ruled that the truth had the same effect as lies.

            Lies cannot be used as defence.

          • Wrong country, dude

          • With that upheld nonsensical ruling that “truth has the same effect as lies” with regard to hatred, the court has undone not only the Foundation of justice but the civilization itself.

            Up is down, right is wrong etc.

            Murder has the same effect as dying of natural causes.

  2. Free speech on social media has the potential to change the world by letting truth known by average citizens challenge the lies and propaganda of the establishment, elites, government and mainstream media.

    This is now under attack by the establishment. Those who bring you propaganda are engaged in a new PR program to combat what they call “fake news” by censoring you and others who seek to use their constitutional right to free speech to inform you.

    All your social media search engines and information platforms are being modified by self proclaimed “fact checkers” to censor us.

    Who are these fact checkers and who is fact checking them?