The 1995 Prodigy Lawsuit That Accidentally Created Section 230

Let me tell you about the best advice a lawyer could have given an online service in 1994: do not clean up. Do not read the messages. Do not delete the spam, do not weed out the abuse, do not so much as glance at the bulletin boards. Be a slob. A messy house, the courts seemed to say, is a safe house. The moment you pick up a broom, you own the dust.

I am not exaggerating by much. Two court decisions, four years apart, produced exactly that incentive, and a pair of congressmen, one from each party, noticed how absurd it was and set out to fix it. What they wrote in 1996 became one of the most debated sentences in American law. You cannot understand the argument over Section 230 today, with all its heat and all its myths, unless you know the story of how a short statute was born from a rule that punished the people who tried to be responsible.

So let us tell it in order, the way a story ought to be told, with the dates and the names, and then ask what the original purpose says about the fight we are having now.

The old rule: publishers and distributors

Start with the common law, which is older than the internet by several centuries. For libel, the law has long drawn a line between two kinds of defendants. A publisher, such as a newspaper, chooses what goes into its pages and is responsible for it. If it prints a libel, it is liable, as though it had said the words itself. A distributor, such as a bookstore or a newsstand, merely passes along what others have produced. A distributor is liable only if it knew, or had reason to know, that the material was defamatory.

That distinction makes sense. The editor who reads every line can be held to every line. The bookseller with ten thousand titles cannot possibly read them all. In 1959 the Supreme Court recognized the point in constitutional terms in Smith v. California, striking down a city ordinance that made a bookseller criminally liable for obscene books on his shelves whether or not he knew what was in them. A rule that made booksellers answerable for every page, the Court reasoned, would lead them to stock only what they were sure about, and the public's access to books would shrink.

The question the courts faced when computer networks arrived was simple to state: when someone posts something defamatory on an online service, is the service a publisher or a distributor?

1991: Cubby v. CompuServe

The first answer came in 1991, in a federal court in New York, in a case called Cubby, Inc. v. CompuServe Inc. CompuServe, one of the early online services, hosted a journalism forum. Someone posted allegedly defamatory remarks about a competing newsletter. The newsletter sued the service.

The court held that CompuServe was a distributor, not a publisher. The service had not reviewed the content before it appeared, had no real editorial control over the forum's daily material and had no knowledge of the particular statements. So it was not liable. The judge likened it to an electronic library or newsstand.

That is a sensible result. But look at the reasoning underneath. CompuServe won because it did not exercise editorial control. The less it looked, the safer it was.

1995: Stratton Oakmont v. Prodigy

Four years later, a state court on Long Island reached the opposite result on very similar facts, and that is what lit the fuse.

The case was Stratton Oakmont, Inc. v. Prodigy Services Co., decided in May 1995 by a New York trial court. Prodigy was an online service that ran bulletin boards, including one called Money Talk, where users discussed investments. An anonymous user posted messages saying that Stratton Oakmont, a securities firm, and its president had committed criminal and fraudulent acts. The firm sued Prodigy for defamation.

Here is the twist. Prodigy had content guidelines. It used software to screen for offensive language. It employed board leaders to enforce its rules and could remove notes that violated them. In other words, Prodigy tried to moderate.

The court held that because Prodigy exercised editorial control, it was a publisher, not a distributor, and so could be held liable for what its users posted. Notice the irony. The very steps Prodigy took to make its service decent were the steps that made it legally responsible for every message on it.

(Anyone who knows the later notoriety of the firm's name may smile at the fact that the plaintiff in the case that created our modern internet law was a brokerage house whose reputation was, shall we say, a matter of continuing public interest. But the legal principle does not depend on who the plaintiff was, and the court was not deciding whether the allegations were true.)

The perverse result

Put the two rulings side by side, and let me follow the logic one step further than its defenders would like.

• Under Cubby, an online service that does not moderate is a distributor, and is safe.
• Under Stratton Oakmont, an online service that does moderate is a publisher, and is exposed.

If you are a service and you read that, the rational move is obvious. Stop moderating. Do not remove the abuse, the harassment, the spam or the fraud, because the instant you do, you are on the hook for everything you missed. The law, which was meant to hold publishers responsible, would reward the service that looked away and punish the service that tried to keep its forums decent. It was a rule that said: ignorance is bliss, and vigilance is expensive.

Anyone who thinks about it for five seconds can see the problem. A parent choosing an online service for a family wants one that removes the worst material. A rule that penalizes such services puts the family-friendly operator at a disadvantage against the free-for-all. That is the opposite of what any sensible policy would want.

1995 and 1996: Congress steps in

Congress was already working on internet legislation, in the form of what became the Telecommunications Act of 1996, a major overhaul of the nation's communications law. Part of the package was the Communications Decency Act, which aimed to restrict indecent material online, a response to widespread public worry about what children might find on the new medium.

Representatives Chris Cox, a Republican from California, and Ron Wyden, a Democrat from Oregon, saw the Stratton Oakmont problem. They proposed an amendment in the House in 1995. Their idea was that online services should be free to remove objectionable material without turning themselves into legal guarantors of everything else, and that users, not services, should bear responsibility for their own words.

The amendment became Section 230 of the Communications Act, codified at 47 U.S.C. § 230. President Clinton signed the Telecommunications Act into law on February 8, 1996.

The heart of it is a single sentence in subsection (c)(1):

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.

Read that carefully, because it is shorter than the arguments about it. A service is not to be treated as the publisher or speaker of information that someone else provided. The person who wrote the defamatory post is the one responsible. The service that hosted it is not treated as having said it.

Subsection (c)(2) adds a companion protection, sometimes called the "Good Samaritan" provision. It shields a service from liability for actions taken in good faith to restrict access to material the service considers "obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable." This is the direct answer to Stratton Oakmont: you may pick up the broom without owning the dust.

The statute also has limits. Subsection (e) provides that the section does not affect federal criminal law, intellectual property law and certain other categories. We will come back to that, because many people claim that Section 230 protects illegal content, and it does not.

1997: the rest of the Communications Decency Act falls

Here is a detail that often gets lost. The part of the 1996 law that grabbed all the headlines at the time was not Section 230. It was the Communications Decency Act's criminal prohibitions on transmitting "indecent" or "patently offensive" material to minors online. Civil liberties groups challenged them immediately.

In June 1997, in Reno v. ACLU, the Supreme Court struck down the indecency provisions of the CDA as violations of the First Amendment. The Court, in an opinion by Justice Stevens, held that the provisions were too broad, restricting a vast amount of protected speech among adults in order to shield minors. The Court treated the internet as deserving the full protection that the First Amendment gives to print, not the lesser protection that applies to broadcasting.

So the headline provisions of the Communications Decency Act were gone. Section 230, the quiet amendment in the same bill, survived. That produced one of the great ironies of modern law: the part of the statute that was meant to be the centerpiece died, and the footnote became the foundation of the commercial internet.

Within months, the courts began applying Section 230 broadly. In Zeran v. America Online, decided by the Fourth Circuit in 1997, the court held that Section 230 barred a suit against the service over anonymous postings, reading the immunity as covering even a service that had been notified of the offending content. That decision set the pattern for the next quarter century.

The Founding anchor: printers, liability and a free press

It is worth setting all this against the longer history. In the colonial period, the printer was the legal target of seditious libel prosecutions, because the printer was the one who could be found, charged and jailed. John Peter Zenger was a printer; see John Peter Zenger and the Jury That Defended American Press Freedom in 1735. When the Sedition Act of 1798 was enforced, printers and editors went to prison; see The Sedition Act of 1798 Jailed a Congressman Who Won Reelection From His Cell. The lesson of that history is that making the intermediary liable for the speech of others is the oldest trick in the censor's book. Go after the person with the press and the speech disappears, because nobody wants the risk.

Section 230, whatever one thinks of its modern application, addresses exactly that danger. If a service had to answer for every word of every user, it would either shut down user speech or filter it to the bone. The First Amendment forbids Congress to abridge "the freedom of speech, or of the press," and a regime of intermediary liability for user speech would have abridged it by proxy. The statute was, in that sense, an act of speech protection, written by a Congress that had just watched courts create the opposite incentive.

What the original purpose says about the debate today

Now for the part where I tell you what I think, and I will be careful to label it opinion.

The statute was meant to encourage moderation, not to forbid it. The purpose was to allow services to remove bad material. Anyone who says that Section 230 requires platforms to be "neutral" has the history backward. The law was written to let them make choices. When critics say that a platform that moderates should lose its immunity, they are asking Congress to restore the exact rule that Stratton Oakmont created and that Section 230 was passed to reverse. You may want that rule for other reasons, but you should know what it is.

The law was written for a different internet. In 1996, a small number of online services hosted bulletin boards and chat rooms. Today a handful of companies host billions of posts, rank them by algorithm and sell advertising against them. A fair-minded person can ask whether a statute written for message boards is the right fit for that business. That is a legitimate question for Congress, and I would welcome a debate on it conducted by people who have read the text.

Reform proposals should be measured against the perverse-incentive problem. Any change to Section 230 should answer the question the original authors answered: does it reward the service that looks away or the service that tries? Proposals to condition immunity on political neutrality, for example, would put a government agency or court in the position of deciding what counts as neutral, which raises the First Amendment concerns I discuss in Moody v. NetChoice and the Unfinished Question of Who Owns the Feed. Proposals to narrow immunity for specific harms may be more workable.

Myths abound. For a point-by-point test of the claims most often heard, such as that moderating forfeits immunity or that 230 protects illegal content, read Publisher or Platform, Eight Myths About Big Tech and the Law. And for the broader question of how the government's relationship to platform moderation should be policed, see Murthy v. Missouri Lost on Standing, Yet the Record It Built Still Stands.

The best objection, answered

The strongest objection to my account says this: whatever Congress intended in 1996, the statute has been read so broadly by the courts that it now shields platforms from almost any responsibility, even for harms the platforms help to cause. A law passed to let services remove bad content has become a license to profit from it. Therefore the original purpose is beside the point, and the law needs reform.

There is real substance here, and I do not dismiss it. Courts have read the immunity expansively, and reasonable people, including some justices, have wondered whether some applications go beyond the text. Congress has already carved out exceptions, for example in 2018 for certain sex-trafficking claims, which shows that the statute is not untouchable.

But notice what the objection actually argues. It does not say the original diagnosis was wrong. It says the cure has been stretched too far. That is an argument for amending the statute precisely, with attention to the consequences, and not for repealing the protection that made a free-speech internet possible. The honest way to have that argument is to start from what the law was designed to fix, and to ask what each proposed change would do to the incentive to moderate and the ability of ordinary people to speak.

The case in brief

1. Cubby v. CompuServe (1991) held an unmoderated service to be a distributor, free of liability; Stratton Oakmont v. Prodigy (1995) held a moderated service to be a publisher, exposed to liability.
2. That combination rewarded services that looked away and punished those that tried to keep forums decent.
3. Representatives Cox and Wyden wrote Section 230, enacted in the Telecommunications Act of 1996, so that services would not be treated as the publisher of others' words and could remove objectionable material without losing protection.
4. The rest of the Communications Decency Act fell in Reno v. ACLU (1997), leaving Section 230 as the surviving core.
5. Any reform should be tested against the original problem: does it encourage or discourage a service from cleaning up?

What you can do: read Section 230 itself on the Cornell Law or Congress.gov websites, all of it, which takes about five minutes, before you take a side in the next argument about it.