Section 230 of the Communications Act is a short federal law with an outsized role. Its central sentence, sometimes called “the 26 words that created the internet,” 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.”
What it does
- Liability shield for user content. A website, app or platform generally can’t be sued for what its users post. The person who wrote a defamatory post can be sued; the platform hosting it usually can’t.
- Protection for moderation. A second provision protects good-faith decisions to remove or restrict content the provider considers objectionable. Platforms can moderate without becoming liable for everything they leave up.
Why it was written
In 1995, a New York court held in Stratton Oakmont v. Prodigy that an online service that moderated its message boards could be treated as the publisher of users’ posts, while an earlier case had found that a service that did no moderation was not. That gave companies an incentive not to moderate at all. Congress responded in 1996 by adding Section 230 to the Communications Decency Act. Courts read it broadly from the start, beginning with Zeran v. America Online (1997).
What it doesn’t cover
- Federal criminal law.
- Intellectual property claims, such as copyright, which has its own notice-and-takedown system.
- Certain sex trafficking claims, under a 2018 amendment known as FOSTA-SESTA.
- Content the platform itself creates or materially develops.
Section 230 also does not stop a platform from setting its own rules; that is protected separately by the First Amendment, as the Supreme Court indicated in Moody v. NetChoice (2024). See is it censorship when Facebook removes your post?
The fights over reform
Critics on the right argue the law lets platforms remove conservative speech without consequence. Critics on the left argue it lets platforms profit from harmful content without responsibility. Proposals have included conditioning protection on political neutrality, carving out algorithmic recommendations, and adding duties of care for children. In Gonzalez v. Google (2023), the Supreme Court was asked whether recommendation algorithms are protected but declined to decide, resolving the case on other grounds.
The AI question
Section 230 protects platforms from liability for information provided by another content provider. When a chatbot generates a false and damaging statement, the text was arguably created by the AI company’s own system, not by a third party. Many legal scholars, and some of the law’s original authors, have said they do not think Section 230 covers generative AI output. Courts have only begun to address the question, and defamation and product-liability suits against AI companies are testing it.