Section 230 Is 26 Words. Almost Everything Online Depends on Them.

What the law says

The operative text of Section 230 is twenty-six words: no provider or user of an interactive computer service shall be treated as the publisher or speaker of information provided by another content provider. In practice this means a platform is generally not liable for what its users post, in the way a newspaper is liable for what it prints.

There is a second half that gets less attention. The law also protects platforms from liability for removing content in good faith. That provision is why Section 230 draws fire from people who want more content removed and from people who want less, sometimes in the same week.

What it never covered

The shield has always had limits. It does not apply to federal criminal law, to intellectual property claims, or, since a 2018 amendment, to conduct related to sex trafficking. Platforms are not immune from prosecution for their own criminal conduct, and the law does not protect content the platform itself created.

That 2018 carve-out is the closest thing to a natural experiment either side has. Supporters of further reform argue it demonstrated the shield can be narrowed without the internet collapsing. Opponents argue it pushed activity to less visible platforms and made some harms harder to police. Researchers disagree, and that disagreement does much of the work in the current debate.

The algorithm question

When the law passed in 1996, platforms mostly displayed posts in the order they arrived. Today ranking systems decide what billions of people see, optimizing for engagement. The central reform argument is that choosing what to amplify is an editorial act, and that a statute written for passive bulletin boards should not automatically cover it.

The counterargument is that ranking is unavoidable at scale. Any feed requires ordering, and treating sorting as publishing would collapse the distinction the law depends on. Several appellate decisions since 2024 have split on exactly this line, which is why the question keeps returning to the Supreme Court's door without being settled.

Why both parties want change and nothing happens

This is one of the few areas where bipartisan appetite for reform has not produced legislation, because the two parties want opposite outcomes. One side generally wants platforms to remove more content and be liable for what they leave up. The other generally wants platforms to remove less and be accountable for what they take down.

A single repeal would move policy in both directions at once, since the law protects hosting and moderation together. That is the structural reason proposals stall: any bill that satisfies one coalition tends to be unacceptable to the other, and the status quo persists by default rather than by agreement.

Where the debate actually splits

The left generally argues that algorithmic amplification is closer to editing than hosting, that immunity has been read broadly enough to cover harms courts would otherwise hear, that narrowing it for paid promotion would leave ordinary speech untouched, and that other democracies impose duties of care without ending user-generated content.

The right generally argues that liability would cause platforms to remove anything remotely risky so lawful speech disappears first, that the rule protects small forums far more than incumbents who can absorb litigation, that repeal could produce less moderation rather than more, and that new standards would be enforced by whichever administration holds power next.

How other countries handle it

The United Kingdom, the European Union and Australia have adopted duties of care requiring large platforms to assess and mitigate specific categories of risk, enforced by regulators rather than through private lawsuits. Supporters of reform treat this as proof that responsibility is possible without ending user-generated content, since none of those countries lost their internet.

Critics note the systems are young, that compliance costs fall hardest on smaller services which cannot staff regulatory teams, and that regulators in several of those jurisdictions have already been accused of pressuring platforms over lawful speech. That last outcome is precisely what the American approach was designed to prevent, which is why the comparison persuades fewer people here than abroad.