--- title: "Section 230" description: "The 26 words that created the internet. Why they made sense in 1996 and what has changed since." url: https://attentionmachine.org/en/laws/section-230/ lang: en updated: 2026-10-25 license: "CC BY-NC-SA 4.0 (https://creativecommons.org/licenses/by-nc-sa/4.0/)" --- # Section 230 In 1996, when 40 million people used the internet[^src-supreme-court-1997], a law was passed. By 2019, 4 billion people were online[^src-international-telecommunicatio-2019]. The law had not changed. ## The 26 words "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." These 26 words, Section 230 of the Communications Decency Act[^src-communications-decency-1996], signed into law on 8 February 1996, are the legal foundation of the modern internet. They gave platforms intermediary immunity: the freedom to host and moderate what users post without being treated as the publisher of it. Civil-liberties groups, the Electronic Frontier Foundation among them, fought the rest of the Communications Decency Act in court and had its censorship provisions struck down[^src-american-civil-1997]; the 26 words survived, and the EFF[^src-electronic-frontier] has defended them since. The intention was reasonable. > **Intermediary immunity** > > The principle that an online service is not legally liable for content its users create, the way a phone company isn't liable for what people say on calls. It is what lets a platform host billions of posts — and moderate them — without being sued over each one. The hard question Section 230 left unanswered: does that immunity also cover the platform's own choice of _which_ posts to amplify? > > **Sources** > > - [Communications Decency Act, Section 230](https://www.law.cornell.edu/uscode/text/47/230) (1996) — Legal Information Institute, Cornell. ## The gap it created Section 230 does not merely protect platforms from liability for what their users post. Courts interpreted it[^src-congressional-research] to protect platforms from liability for their own algorithmic recommendations — for actively choosing to show you content that plaintiffs and some studies link to radicalisation, depression, or harm, though whether it causes them is contested. A platform that amplifies harmful content through its recommendation engine is, under Section 230's interpretation, no more liable than one that merely hosts it passively. When the law was written, 40 million people used the internet[^src-supreme-court-1997]. By 2019, more than 4 billion were online[^src-international-telecommunicatio-2019]. Platforms had built multi-billion-dollar businesses on algorithmic amplification. The 26-word immunity, written for a different world, applied to this one. ## What is changing US courts are now finding that Section 230 does not protect platforms from liability for some of their own design choices — parental controls, age checks, the barriers to deleting an account, the failure to warn — as distinct from the content itself. The feed's own features are the contested middle: in 2023 the federal judge in the youth-harm litigation held claims against the endless feed, the recommendation algorithms and notification timing barred[^src-dla-2024], while in 2024 the Third Circuit let a suit over TikTok's recommendations go ahead[^src-court-appeals-2024]. The March 2026 Los Angeles verdict against Meta and Google's YouTube[^src-cnbc-2026] was won partly on this argument: not "you hosted bad content" but "you designed a product you knew was harmful." The immunity is not absolute. The design argument is gaining force. **How we know — the content-versus-design distinction courts are drawing** The legal shift turns on a single distinction. Section 230 immunises a platform for _third-party content_[^src-communications-decency-1996] — what users say. It was never obvious that it also immunises the platform's _own product decisions_ — infinite scroll, autoplay, recommendation, notification timing. Recent rulings (the November 2023 federal decision in the youth-harm litigation[^src-dla-2024], allowing failure-to-warn claims and design claims about parental controls, age checks and account deletion, while holding the endless-feed, recommendation and notification-timing claims barred; the Third Circuit allowing a suit over TikTok's algorithm[^src-court-appeals-2024]; the March 2026 verdicts[^src-cnbc-2026]) accept that at least some design is the platform's own conduct, not someone else's speech, and so falls outside the 26 words. **What's still unsettled.** Where exactly the line sits between "a recommendation is editorial speech (protected)" and "a recommendation is a product feature (liable)" is being litigated case by case, and appeals are ongoing[^src-cnbc-2026]. The direction is consistent; the precise boundary is not yet fixed law. ## Sources [^src-communications-decency-1996]: [Communications Decency Act, Section 230](https://www.law.cornell.edu/uscode/text/47/230) (1996) — Legal Information Institute, Cornell. [^src-american-civil-1997]: American Civil Liberties Union, [_Reno v. ACLU_](https://www.aclu.org/cases/reno-v-aclu-challenge-censorship-provisions-communications-decency-act) (1997) — the challenge to the CDA's censorship provisions, with the EFF among the plaintiffs; Section 230 was untouched. [^src-supreme-court-1997]: Supreme Court of the United States (1997), [_Reno v. American Civil Liberties Union_, 521 U.S. 844](https://www.law.cornell.edu/supremecourt/text/521/844) — Legal Information Institute, Cornell: "About 40 million people used the Internet at the time of trial" (1996). [^src-international-telecommunicatio-2019]: International Telecommunication Union (2019), [_Measuring digital development: Facts and figures 2019_](https://www.itu.int/en/ITU-D/Statistics/Documents/facts/FactsFigures2019.pdf) — "An estimated 4.1 billion people are using the Internet in 2019" (ITU estimate). [^src-electronic-frontier]: [Electronic Frontier Foundation — Section 230](https://www.eff.org/issues/cda230). [^src-congressional-research]: [US Congressional Research Service, _Section 230: An Overview_](https://crsreports.congress.gov/product/pdf/R/R46751) (R46751). [^src-cnbc-2026]: CNBC (10 June 2026), [_Google and Meta denied new trial in youth social media addiction case_](https://www.cnbc.com/2026/06/10/google-and-meta-denied-new-trial-in-youth-social-media-addiction-case.html); Fortune (15 July 2026), [_YouTube appeals verdict, argues it isn't a social media platform_](https://fortune.com/2026/07/15/youtube-appeals-social-media-addiction-verdict/) — Meta filed its notice of appeal days earlier. [^src-dla-2024]: DLA Piper (January 2024), [_Court addresses social media liability in adolescent addiction litigation_](https://www.dlapiper.com/en/insights/publications/2024/01/navigating-the-digital-dilemma-court-addresses-social-media-liability-in-adolescent-addiction) — summary of the 14 November 2023 ruling on Section 230 and the First Amendment; the order itself: US District Court for the Northern District of California (14 November 2023), [_In re: Social Media Adolescent Addiction/Personal Injury Products Liability Litigation_, No. 4:22-md-03047-YGR, Dkt. 430](https://storage.courtlistener.com/recap/gov.uscourts.cand.401490/gov.uscourts.cand.401490.430.0.pdf) — claims over parental controls, age verification, barriers to deleting an account and failure to warn not barred; claims over the feed's lack of a "beginning and end", the "use of algorithms to promote addictive engagement" and the timing and clustering of notifications barred. [^src-court-appeals-2024]: US Court of Appeals for the Third Circuit (27 August 2024), [_Anderson v. TikTok_](https://www2.ca3.uscourts.gov/opinarch/223061p.pdf). [GDPR →](https://attentionmachine.org/en/laws/gdpr/)\ [Courts & Liability →](https://attentionmachine.org/en/laws/courts/) --- Some content on this site was created with the assistance of AI tools and may contain errors. [How we use AI →](https://attentionmachine.org/en/about/how-we-use-ai/)