What is Section 230?
Section 230 of the Communications Decency Act, passed in 1996, is the sentence that decides who answers for the feed: twenty-six words holding that a platform is not the publisher of what its users post. A companion clause protects whatever the platform chooses to take down. A federal appeals court read the shield broadly within a year, and in 2024 the Supreme Court closed the frame from the other side: the feed's moderation and ranking choices are editorial judgments, protected by the First Amendment like a newspaper's. Together the documents describe a gatekeeper with an editor's power over the record and almost none of an editor's exposure to a courtroom.
1996: the twenty-six words
Congress opened the section by stating its theory of the new medium, and it is worth reading against everything the rest of this act documents:
Primary source 01 / 09
The Internet and other interactive computer services offer a forum for a true diversity of political discourse, unique opportunities for cultural development, and myriad avenues for intellectual activity.
Then it wrote the sentence the modern feed is built on:
Primary source 02 / 09
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.
Count them: twenty-six words. The next clause protects the opposite power, removal, and sets the standard for what may be removed as the platform's own opinion:
Primary source 03 / 09
any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected
The operative words are considers to be and otherwise objectionable. What a platform may take down is measured by the platform's own judgment, and the protection holds, the statute says, whether or not the material was constitutionally protected speech.
1997: immunity, read as broadly as possible
The first major test was ugly. Six days after the Oklahoma City bombing, an anonymous AOL user advertised T-shirts with slogans mocking the attack and listed the home phone number of a Seattle man named Kenneth Zeran, who had nothing to do with any of it. The calls included death threats. More postings followed, and Zeran sued AOL, arguing it unreasonably delayed removing the messages. The Fourth Circuit held that the new statute barred the suit entirely:
Primary source 04 / 09
creates a federal immunity to any cause of action that would make service providers liable for information originating with a third-party user of the service.
And the court named what the immunity covers, in the vocabulary of this encyclopedia:
Primary source 05 / 09
lawsuits seeking to hold a service provider liable for its exercise of a publisher’s traditional editorial functions—such as deciding whether to publish, withdraw, postpone or alter content—are barred.
Deciding whether to publish, withdraw, postpone or alter content is a working definition of an editor's job. From 1997 on, American law treated the platform's handling of user content as editing that could not be sued over, whichever way the edit cut: leave it up, take it down, never look at it at all.
2024: the feed is the editorial page
Section 230 shields the platform from liability for its users' speech. The harder question waited until the feed became the main stage of politics: is the moderation itself, the ranking, labeling, and removal, the platform's own speech? Texas and Florida forced the answer. In 2021 both states passed laws restricting how large platforms could moderate; the record before the Supreme Court reflected that Texas officials passed theirs because they thought the feeds skewed against conservative voices. The platforms' trade association sued, and in Moody v. NetChoice, Justice Elena Kagan, writing for the Court, described what a feed is:
Primary source 06 / 09
In constructing certain feeds, those platforms make choices about what third-party speech to display and how to display it. They include and exclude, organize and prioritize—and in making millions of those decisions each day, produce their own distinctive compilations of expression.
Include and exclude, organize and prioritize: the Court described an algorithmic feed the way act two of this encyclopedia describes a newsroom. Then it settled the constitutional status of those choices:
Primary source 07 / 09
At least on the current record, the editorial judgments influencing the content of those feeds are, contrary to the Fifth Circuit’s view, protected expressive activity. And Texas may not interfere with those judgments simply because it would prefer a different mix of messages.
The opinion then reached the question this whole encyclopedia circles, whether anyone can order a feed to be fair, and answered it for governments in one line:
Primary source 08 / 09
it is no job for government to decide what counts as the right balance of private expression—to “un-bias” what it thinks biased, rather than to leave such judgments to speakers and their audiences.
The Court vacated both appeals-court decisions and sent the cases back, because the challenges attacked the laws as a whole and the lower courts had not weighed their full range of applications. But it left the principle standing:
Primary source 09 / 09
But that work must be done consistent with the First Amendment, which does not go on leave when social media are involved.
An editor no one can sue
Put the three documents together and the shape of the modern gatekeeper appears. By statute, the platform is not the publisher of what you post, so it cannot be sued for carrying it. By statute again, its good-faith removals are protected on its own definition of objectionable. And by constitutional holding, its ranking and moderation are its own protected speech. Compare the broadcast era: there, a unanimous Supreme Court held the audience's rights paramount over the broadcaster's and let the government order balance onto the airwaves. For the feed, the polarity is reversed. The mix of messages is the platform's protected expression, and the government may not “un-bias” it.
That is the legal answer to the question this act keeps raising. The system Frances Haugen described from inside, the one Twitter's own study measured a tilt in, is in law an editor: lawful in what it carries, protected in what it cuts, and under no obligation to show anyone the difference. The remedy, if there is one, is not coming from a courtroom; it is the reader's own habit of comparison. The final chapter reaches the machines that generate the record itself.