Publisher or Platform, Eight Myths About Big Tech and the Law

There is a sentence you have heard at a dinner table, in a radio call, in a congressional hearing and, I am quite sure, in the comments under a video about lawn care: "They are a publisher, not a platform, so they lose their protection." Say it with confidence and people nod. Ask the speaker which section of the statute contains the words "publisher, not a platform" and watch the nodding stop.

Folks, I like a good argument as much as anyone, and the argument over big tech is worth having. The companies have enormous power, they make mistakes and they deserve scrutiny. But a debate conducted on myths produces bad laws, and conservatives of all people should insist on reading the text before we legislate. So let us do it. Eight claims, each stated fairly, each tested against the statute and the cases, each given a verdict.

A word on the method. The text of Section 230 of the Communications Act, 47 U.S.C. § 230, is short enough to read over lunch. The key sentence says: "No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider." For the story of how that sentence came to be, read The 1995 Prodigy Lawsuit That Accidentally Created Section 230. Armed with it, on to the myths.

Myth 1: "If a platform moderates content, it becomes a publisher and loses Section 230"

The claim. A company that removes posts or bans users is exercising editorial control, so it is a publisher, and publishers are not protected.

The test. Read the statute. Nothing in Section 230 conditions protection on the platform being neutral or declining to moderate. Subsection (c)(1) protects a provider from being treated as the publisher of content provided by others. Subsection (c)(2) separately protects a provider's good-faith actions to restrict access to material it considers "obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable." Congress wrote (c)(2) precisely so that moderating would not create liability.

And the history points the same way. The statute exists because a 1995 New York ruling treated a service that moderated as a publisher, and Congress wanted to reverse that. To say that moderation forfeits the protection is to say the statute means the opposite of what it was passed to accomplish.

The verdict. False. The text protects moderation. If you want a rule that moderation costs a company its immunity, you are asking Congress to repeal what it enacted, which is a respectable position but a different one.

Myth 2: "The law requires platforms to be neutral"

The claim. Section 230 was a deal: immunity in exchange for political neutrality.

The test. Look for the word "neutral" in the text. It is not there. The statute's findings, in subsection (a)(3), describe interactive services as offering a forum for "a true diversity of political discourse." That is a statement of the environment Congress hoped for. Nothing in the text makes it a condition of immunity.

There is also a constitutional problem with the claim. A rule requiring neutrality as a condition of a government benefit would make a federal agency or court the judge of what is neutral. That raises the First Amendment issues discussed in Moody v. NetChoice and the Unfinished Question of Who Owns the Feed, where the Supreme Court said in 2024 that curating a feed is protected editorial activity.

The verdict. False as a description of the law. As a policy proposal, it is debatable, but a conservative should ask who the neutrality referee would be and who would appoint him.

Myth 3: "The First Amendment applies to private platforms, so banning a user is unconstitutional censorship"

The claim. The First Amendment protects free speech, so when a platform bans someone, it violates the Constitution.

The test. Read the first words of the First Amendment: "Congress shall make no law." The amendment restrains government. Through the Fourteenth Amendment it restrains the states. It does not restrain a private company. This is the state-action doctrine, and the Supreme Court restated it in 2019 in Manhattan Community Access Corp. v. Halleck. In that case, the Court, by a vote of five to four in an opinion by Justice Kavanaugh, held that a private nonprofit running public-access cable channels was not a state actor and so was not bound by the First Amendment. The opinion explained that merely hosting speech by others is not a traditional, exclusive public function, and does not transform a private entity into a government actor.

The principle cuts both ways, and conservatives should relish that. The same rule means a conservative bookstore need not stock a book it despises, and a church need not host a speaker who mocks its faith. The First Amendment protects your right against the government, and it protects the private editor's right to choose from the government.

The verdict. False in the sense claimed. A platform's ban may be unwise, unfair or a betrayal of its own stated values, and you may criticize it. It is not, by itself, a First Amendment violation. The constitutional line is crossed when the government is behind the removal; see NRA v. Vullo and the Unanimous Rule Against Government Jawboning for that rule.

Myth 4: "Section 230 protects illegal content"

The claim. Section 230 lets platforms host anything, including criminal material, without consequence.

The test. Section 230(e) lists what the section does not do. It does not impair the enforcement of federal criminal law. It does not limit intellectual property law. It does not affect certain communications privacy laws. And, since a 2018 amendment, it does not bar certain claims related to sex trafficking. The immunity is from being treated as the publisher or speaker of someone else's content in civil claims such as defamation, not from the federal criminal code.

Two refinements. First, because the criminal-law carve-out refers to federal law, state criminal statutes are in a more complicated position, and courts have addressed that in different ways. Second, Section 230 does not protect a company from liability for its own unlawful conduct, as distinct from its role as host of others' speech.

The verdict. Mostly false. The statute has an explicit federal criminal-law exception and other carve-outs. Where a platform is accused of violating federal criminal law, Section 230 is not a shield.

Myth 5: "Repeal Section 230 and the censorship ends"

The claim. If Congress repealed the law, platforms would have to stop silencing conservatives.

The test. Follow the logic one step further than its fans do. Without Section 230, a platform would return to the legal world of 1995. A service that screens content might be treated as a publisher and held liable for what users post. A service that does not screen might avoid that, but it would still face lawsuits and the cost of defending them. What does a rational company do? It does one of two things. It moderates more, removing anything that might draw a lawsuit, because the cost of a mistaken "leave it up" is a lawsuit and the cost of a mistaken "take it down" is nothing. Or it stops hosting user content altogether.

Neither outcome helps a person who wants to post controversial opinions. A platform facing liability for every post will err on the side of removal. The loudest voices lose first, and small sites and individual bloggers with comment sections, who cannot afford a litigation department, lose with them. Meanwhile the giants, who can afford lawyers and filters, would hold their positions.

On top of that, repeal would not change the constitutional point from Myth 3 or from Moody. The platforms would still have First Amendment rights to curate. Repeal would simply remove a procedural protection from lawsuits.

The verdict. False. Repeal is more likely to produce more removal than less. A serious reform would be targeted and would be tested against that perverse incentive.

Myth 6: "Section 230 is a gift to Big Tech"

The claim. The law is a special privilege for a handful of giant companies.

The test. The statute protects any "provider or user of an interactive computer service." That includes a local news site with a comment section, a hobbyist's forum, a church's online prayer board, a small business's review page and, notably, an ordinary individual who forwards or shares someone else's post. Section 230 was written in 1996, long before the giants existed, when the typical beneficiary was a small service with a bulletin board. Many of today's largest companies did not exist.

That does not mean the giants do not benefit. They do, enormously. But a law that protects everyone who hosts or shares another person's words is not a bespoke subsidy for a few companies. It is a general rule about who is liable for whose speech. Narrowing it may hit the little guy harder than the giant.

The verdict. Misleading. The giants benefit, but so does every small site and every user. A reform designed to punish the biggest companies could end up burdening the smallest.

Myth 7: "The Supreme Court struck down Section 230" or "Section 230 was part of a law that was struck down"

The claim. I have heard both versions: that the Court killed the law long ago, and, from the other direction, that the law is a settled constitutional pillar blessed by the justices.

The test. In Reno v. ACLU (1997), the Court struck down the Communications Decency Act's provisions criminalizing the transmission of "indecent" or "patently offensive" material to minors. Section 230 was in the same bill, as part of the Telecommunications Act of 1996, but it was not the target of that challenge and it survived. And the Court has never held Section 230 unconstitutional.

The opposite claim is also overstated. The Supreme Court has not issued a major merits ruling interpreting the scope of Section 230. In 2023, in Gonzalez v. Google, the Court took a case that raised questions about whether the statute protects algorithmic recommendations, and then declined to decide those questions, sending the case back in light of a companion decision. So the broad readings of Section 230 that govern today come mainly from the lower courts, beginning with Zeran v. America Online (Fourth Circuit, 1997), not from a Supreme Court holding.

The verdict. Both versions are false. The statute stands, it has not been struck down, and the Supreme Court has so far left its boundaries to Congress and the lower courts. That is a reason for legislators to speak clearly.

Myth 8: "Platforms are immune for everything they do"

The claim. Section 230 means a platform can never be held responsible for anything that happens on its service.

The test. The immunity covers information "provided by another information content provider." Where a platform itself creates or helps to develop the unlawful content, it is an information content provider, and the protection does not apply to that portion. A leading example is Fair Housing Council v. Roommates.com, a 2008 decision of the Ninth Circuit sitting en banc, which held that a housing site that required users to answer discriminatory questions as a condition of using the service had helped develop the content and was not immune for it.

Also, the immunity does not shield a company from claims about its own conduct, such as breaking its own promises or violating other laws that do not treat it as the publisher of someone else's words. And it does not shield it from the federal criminal and intellectual property exceptions described in Myth 4.

The verdict. False. The immunity is real but bounded, and it applies to liability as the publisher or speaker of third-party content.

What the myths have in common

Look at the list as a whole. Each myth takes something true and stretches it past breaking. It is true that platforms have power. It is true that moderation has sometimes looked one-sided. It is true that Section 230 is broadly read. From those grains we assemble a story in which a single statute is the cause of everything we dislike, and a single repeal would be the cure.

The truth is less satisfying and more useful. The legal protections here are modest and specific. The real problems, concentration of power, opaque enforcement and the entanglement of government and platforms, call for different tools. Antitrust enforcement for monopoly conduct. Transparency about how rules are applied. Limits on government pressure of the kind documented in Murthy v. Missouri Lost on Standing, Yet the Record It Built Still Stands and addressed in The Twitter Files Showed a Process, Not a Plot, and That Is Worse. Competition from new services. None of these requires believing a myth.

Notice also that the conservative instinct should lean toward the rule of law and away from the slogan. A reform written in anger at one company will be applied to every company, and to every conservative blogger, church and small business with a comment section. The Founders designed a system in which statutes are drafted carefully, because they last longer than the grievances that produced them. For the text-first habit, the constitution page is a good place to practice, and the glossary has the legal terms used above.

The case in brief

1. Moderation does not forfeit Section 230. The statute was written to protect it, and there is no "publisher or platform" test in the text.
2. The First Amendment restrains government, not private platforms, as Halleck (2019) confirms, though government pressure on platforms is a constitutional problem of its own.
3. Section 230 does not protect illegal content in the sense claimed: it has carve-outs for federal criminal law, intellectual property and certain trafficking claims, and it does not cover a platform's own content.
4. Repealing it would more likely increase removal than reduce it, and it protects small sites and ordinary users as well as giants.
5. The Supreme Court has not struck down or definitively interpreted Section 230, so reform belongs to Congress, and it should be drafted from the text.

What you can do: read Section 230(c) and (e) in full, then ask your representative which specific change to that text they support and what it would do to small websites.