Free Speech Articles: Court Rulings, Book Bans, and AI

Free speech in the United States is the right, grounded in the First Amendment, to express ideas without government punishment or prior restraint, subject to a short list of narrow exceptions the Supreme Court has carved out over a century. The amendment binds federal, state, and local government and all three branches, but not private employers, platforms, or publishers. Within that framework, 2023 through 2026 have produced some of the most consequential rulings and political fights over expression in a generation.

What the First Amendment Actually Protects

The constitutional text addresses only Congress, but the Supreme Court long ago extended its protections against every level of government. It does not restrict private individuals or organizations. A newspaper, a social media company, or a university that is not a state actor can set its own rules about what it will publish or host without triggering the First Amendment.

“Speech” and “press” are read broadly. They cover writing, broadcasting, internet communication, and symbolic expression such as flag burning and the wearing of armbands. Entertainment, vulgarity, blasphemy, and speech commonly labeled “hate speech” are all protected. The guiding principle is that content-based restrictions, where the government targets a particular message, are presumptively unconstitutional because the government is not trusted to decide which ideas citizens may hear.¹

Content-neutral regulation is different. The government can impose reasonable time, place, and manner rules — noise ordinances, permits for parades, traffic controls at demonstrations — as long as they apply equally to all speakers regardless of viewpoint.

What Speech the Government Can Restrict

A small set of categories sits outside the First Amendment’s protection. Within these, the government has far more room to regulate or punish:

Incitement to imminent lawless action, under the standard set in Brandenburg v. Ohio (1969).
True threats of violence, as refined in Counterman v. Colorado (2023).
Fighting words likely to provoke an immediate violent reaction.
Obscenity, defined by the three-part test in Miller v. California (1973).
Child pornography involving actual minors.
Defamation, fraud, and perjury.
Speech integral to criminal conduct.²

Hate speech, as such, is not on this list. It is restricted only when it crosses into one of the categories above, most often incitement or a true threat. That is a sharp departure from most other democracies.

The Cases That Built the Modern Doctrine

A handful of Supreme Court decisions do most of the work in American free speech law:

Schenck v. United States (1919) introduced the “clear and present danger” test for restricting speech that poses an immediate risk of serious harm.³
Gitlow v. New York (1925) applied the First Amendment to state governments through the Fourteenth Amendment’s due process clause.⁴
New York Times Co. v. Sullivan (1964) required public officials suing for defamation to prove “actual malice,” meaning knowledge of falsity or reckless disregard for the truth.³
Brandenburg v. Ohio (1969) replaced the “clear and present danger” test with a stricter rule: advocacy of force or law-breaking is protected unless it is directed at inciting imminent lawless action and is likely to produce it.⁵
Tinker v. Des Moines (1969) held that students “do not shed their constitutional rights at the schoolhouse gate” and may express political views unless they materially disrupt school operations.⁵
Texas v. Johnson (1989) protected flag burning as symbolic political speech.⁵
Citizens United v. FEC (2010) struck down limits on corporate independent campaign expenditures, holding that the government may not suppress political speech based on a speaker’s corporate identity.³

What the Supreme Court Has Done Recently

Between 2023 and 2025 the Court decided a run of cases that have redrawn parts of this framework. Several are already changing how prosecutors, platforms, and states operate.

Counterman v. Colorado (2023)

In a 7–2 decision written by Justice Kagan, the Court held that prosecutors in true-threats cases must prove the defendant subjectively understood the threatening nature of their statements. Recklessness, meaning conscious disregard of a substantial risk that the message would be read as threatening violence, is enough.⁶ The decision replaced a purely objective “reasonable person” standard and gave what the Court called “breathing room” against self-censorship.⁷ Justices Barrett and Thomas dissented, favoring an objective test.

303 Creative LLC v. Elenis (2023)

The Court held that the First Amendment prevents a state from forcing a website designer to create expressive designs conveying messages the designer disagrees with. Doing so would be impermissible compelled speech.³

Moody v. NetChoice (2024)

Florida and Texas had passed laws restricting social media platforms from removing certain political content. The Court unanimously vacated the lower-court decisions and sent the cases back for proper analysis. Justice Kagan’s majority opinion affirmed that platforms’ content-moderation decisions are protected editorial judgments under the First Amendment, analogous to a newspaper’s right to choose what to publish, and rejected the argument that platforms are mere conduits that states can regulate for “viewpoint neutrality.”⁸ Enforcement of both state laws remains paused while the challenges continue below.⁹

Murthy v. Missouri (2024)

This case asked whether White House officials, the Surgeon General, the CDC, the FBI, and CISA violated the First Amendment by pressuring social media companies to suppress speech about COVID-19 and the 2020 election. In a 6–3 decision, the Court reversed the Fifth Circuit and vacated a preliminary injunction against the government, holding that no plaintiff had Article III standing. Justice Barrett’s majority opinion found the record lacked specific evidence tying platform moderation to government coercion rather than to the platforms’ own policies.¹⁰ The Court did not resolve when government communication to a private platform crosses from persuasion into coercion.¹¹

The underlying dispute did eventually settle. In March 2026 the parties entered a consent decree imposing a 10-year court-enforceable injunction. The Surgeon General, CDC, and CISA are permanently barred from threatening Facebook, Instagram, X, LinkedIn, and YouTube with legal, regulatory, or economic punishment to induce removal or suppression of the named plaintiffs’ protected speech. The government may still label posts as inaccurate or contrary to its views, so long as it does not couple those statements with threats.¹²¹³

TikTok Inc. v. Garland (2025)

On January 17, 2025, the Court upheld the Protecting Americans from Foreign Adversary Controlled Applications Act, which requires TikTok’s Chinese parent ByteDance to sell the platform or face a U.S. ban. Applying intermediate scrutiny, the Court found the law facially content-neutral because it targets a foreign adversary’s control over the platform, not the content of anyone’s speech. Preventing China from collecting sensitive data on roughly 170 million American users was deemed “decidedly content agnostic.”¹⁴ The Court stressed the “inherent narrowness” of its holding.¹⁵

Free Speech Coalition v. Paxton (2025)

On June 27, 2025, in a 6–3 decision, the Court upheld a Texas law requiring commercial websites hosting sexually explicit content to verify visitors’ ages. Writing for the majority, Justice Thomas held the law subject to intermediate scrutiny because it only “incidentally burdens” adult speech, and that states retain traditional power to prevent minors from accessing material obscene as to them. The majority drew parallels to age-gating for alcohol and firearms.¹⁶ FIRE called the decision one that reversed “decades of precedent that protected the free speech rights of adults.”¹⁷

Platforms, Section 230, and Government Pressure

Section 230 of the Communications Decency Act, enacted in 1996, is the legal backbone of online content moderation. Its core provision states that 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.”¹⁸ The statute lets platforms moderate as they see fit and shields them from most civil liability for user content. It does not cover federal criminal violations, intellectual property claims, or content the platform itself creates.

The Electronic Frontier Foundation argues that without this protection, online intermediaries would either filter aggressively or stop hosting user content.¹⁸ Legal scholars have made a parallel First Amendment argument: something like Section 230 may be constitutionally necessary to prevent “collateral censorship,” where intermediaries over-remove user speech to avoid litigation costs.¹⁹

Beyond Section 230 itself, the question of when government pressure on platforms becomes unconstitutional coercion is actively contested. The JAWBONE Act, introduced in June 2026 by Senators Ted Cruz and Ron Wyden, would create a federal cause of action against government officials who coerce broadcasters, social media platforms, or AI providers into censoring lawful speech. It would set up a transparency portal for government-platform communications and provide remedies including compensation and attorney fees.²⁰²¹ The ACLU, FIRE, and eight other organizations have endorsed it.

Speech on Campus

American colleges have become one of the most active enforcement sites for speech disputes. FIRE and College Pulse surveyed over 68,000 undergraduates at 257 institutions for their 2026 College Free Speech Rankings. Of those schools, 166 received a failing grade and only 11 earned a C or higher. Claremont McKenna College ranked first; Barnard, Columbia, and Indiana University sat at the bottom.²²

For the first time, a majority of surveyed students opposed allowing controversial speakers from either the political left or right on campus. A record one in three said using violence to stop campus speech is acceptable. Fifty-three percent cited the Israeli-Palestinian conflict as the most difficult topic to discuss openly.²³ FIRE recorded 273 efforts to suppress student speech in 2025, surpassing its previous high of 252 in 2020.²⁴

Several incidents have ended up in federal court. At Indiana University, a federal judge found the university’s 2024 protest policy unconstitutional and ordered student reprimands expunged. At Ohio State, a judge ruled that the university violated a student’s due process rights and retaliated against protected speech when it expelled him without a hearing after controversial comments. At Columbia, four students were suspended for hosting a panel and three have sued; journalists at Barnard and Columbia also faced suspensions following pro-Palestinian protests.²⁴

State Protest Laws

Texas enacted Senate Bill 2972, effective September 1, 2025, banning “expressive activity” on university campuses between 10 p.m. and 8 a.m., limiting amplified sound during class hours, prohibiting building encampments, and requiring students and employees to produce identification on request.²⁵ FIRE sued the University of Texas System on September 3, 2025. On October 14, 2025, U.S. District Judge David Ezra temporarily blocked parts of the law, finding the plaintiffs likely to succeed on First Amendment grounds. The UT System appealed on November 4, 2025, and the appeal remains pending.²⁶

Nationally, since 2023 at least 29 states have considered over 135 bills aimed at reducing university autonomy, restricting academic freedom, or limiting expression, with at least 27 enacted.²⁵ The ICNL US Protest Law Tracker shows that 45 states have considered 384 bills affecting protest rights since 2017, with 57 enacted and 43 pending as of mid-2026. Federal proposals in the pipeline include bills that would add “riot” offenses to RICO statutes, bar federal student aid for students convicted of protest-related offenses, create federal felonies for blocking highways, and mandate deportation of non-citizens convicted of protest-related crimes.²⁷

Deportations Tied to Political Speech

Starting in early 2025, the Trump administration began targeting international students and scholars for deportation, often citing their political speech regarding Israel and Palestine under the “foreign policy ground” of the Immigration and Nationality Act. The most prominent case involves Mahmoud Khalil, a Palestinian lawful permanent resident and Columbia University graduate who served as a lead negotiator during pro-Palestine campus protests. Khalil was arrested by federal immigration agents on March 8, 2025, and his attorneys filed a habeas petition arguing retaliation for protected speech.²⁸ In January 2026, a federal judge ruled that the administration’s broader policy of targeting international students and faculty for their speech was an “unconstitutional conspiracy.”²⁴ A Third Circuit decision in January 2026 vacated a lower-court ruling in Khalil’s case on jurisdictional grounds, and his legal team has said it will seek Supreme Court review.²⁹

Book Bans and Classroom Censorship

Restrictions on what students can read and learn have accelerated. In 2022, the United States saw the highest volume of attempted book bans in 20 years, with more than 1,000 recorded demands to censor library books, many targeting titles by or about Black and LGBTQ individuals. In 2023, over 100 classroom censorship bills were introduced nationwide and 32 were signed into law.³⁰ The Department of Defense has banned more than 500 titles in its military-operated schools, citing concerns about materials promoting “woke” ideologies regarding race, gender, and LGBTQ issues.³¹

The ACLU is litigating several of these cases. E.K. v. Department of Defense Education Activity, filed in 2025 by 12 students, challenges the book removals and curriculum changes in military schools following executive orders banning “divisive concepts.” In BERT v. O’Connor, a challenge to Oklahoma’s HB 1775, the Oklahoma Supreme Court ruled in 2025 that the law does not apply to higher education; the K-12 portion remains on appeal. In Pernell v. Lamb, a trial court blocked enforcement of Florida’s “Stop W.O.K.E. Act” in higher education, citing First and Fourteenth Amendment violations, and the state has appealed.³²

AI-Generated Content

AI programs themselves have no First Amendment rights. They are tools, not speakers. The content they produce, however, is generally treated as speech entitled to the same protections as human-generated expression, subject to the usual unprotected categories like defamation, fraud, and obscenity.³³

In Walters v. OpenAI, decided by a Georgia state court on May 19, 2025, a radio host sued OpenAI after ChatGPT generated false claims of embezzlement about him. The court granted summary judgment for OpenAI, finding that a reasonable user, aware of ChatGPT’s tendency to “hallucinate” and its prominent disclaimers, would not interpret the output as asserting actual facts. The court declined to impose what it characterized as a strict-liability standard on AI developers.³⁴ The practical implication is that developers who implement reasonable safeguards and adequate warnings are not automatically liable for AI-generated falsehoods.

On the legislative side, the White House released a National Policy Framework for Artificial Intelligence in March 2026 recommending that Congress prevent the government from “coercing technology providers, including AI providers, to ban, compel, or alter content based on partisan or ideological agendas.” The framework urged protections for parody, satire, and news reporting within any federal digital-replica law, and recommended preempting state AI regulations.³⁵ The Take It Down Act, signed into law on May 19, 2025, criminalized nonconsensual intimate imagery including AI-generated deepfakes and requires platforms to remove flagged content within 48 hours. The first conviction under the act occurred on April 9, 2026.³⁶ Critics, including the Cato Institute, have raised free speech concerns about the law’s lack of a counter-notice mechanism and the incentive it creates for platforms to over-remove lawful content.³⁷

Executive Action in 2025

On January 20, 2025, his first day back in office, President Trump signed an executive order titled “Restoring Freedom of Speech and Ending Federal Censorship.” The order declares that the First Amendment “enshrines the right of the American people to speak freely in the public square without Government interference” and prohibits the federal government from using taxpayer resources to “unconstitutionally abridge the free speech of any American citizen.” It condemns efforts to coerce social media companies into moderating or suppressing disfavored speech and directs the Attorney General to investigate such activities from the prior four years.³⁸ The order carries standard disclaimers that it creates no enforceable legal rights.

Critics have noted a tension between the order’s stated commitment and the administration’s other actions, including pressuring the FCC to scrutinize broadcast content, threatening broadcast licenses, targeting international students for political speech, and dismantling government-funded media outlets. A June 2026 analysis by the Carnegie Endowment for International Peace described American free speech as a site of “toxic political polarization” where both major political factions invoke free speech principles selectively.³⁹

How the U.S. Compares Abroad

The American approach is among the most permissive in the world. Hate speech is protected unless it both intends to incite violence and creates a likelihood of imminent violence, and the government has no legal obligation to require platforms to remove offensive third-party content.⁴⁰

The European model, rooted in Article 10 of the European Convention on Human Rights, treats free expression as a right that “carries with it duties and responsibilities.” EU law requires criminalization of hate speech that “publicly incites to violence or hatred” based on race, color, religion, descent, or national or ethnic origin, and many member states go further.⁴⁰ The Digital Services Act requires “Very Large Online Platforms” with over 45 million monthly EU users to identify and mitigate “systemic risks” including the spread of illegal content and harms to electoral processes, with fines up to 6% of global revenue for noncompliance.⁴¹ Critics warn of a “Brussels effect,” where platforms apply EU-mandated standards globally rather than maintaining region-specific rules.⁴²

At the international level, ICCPR Article 19 guarantees the right to “seek, receive and impart information and ideas of all kinds,” but permits restrictions that are provided by law, serve a legitimate purpose, and are necessary and proportionate. ICCPR Article 20 requires states to prohibit advocacy of national, racial, or religious hatred that constitutes incitement to discrimination, hostility, or violence.⁴³ The gap with U.S. law is widest on hate speech and blasphemy, where American courts protect expression that many other jurisdictions restrict.⁴⁴

Press Freedom

The 2026 Reporters Without Borders World Press Freedom Index reported that global press freedom has hit a 25-year low. For the first time, over half of the 180 countries surveyed fell into the “difficult” or “very serious” categories, up from 13.7% in 2002. Only about 1% of the world’s population now lives in a country with a “good” media environment. The legal framework indicator saw the steepest decline of all five index metrics.⁴⁵

The United States dropped seven places to 64th, its lowest ranking ever. RSF attributed the decline to what it described as a “coordinated war on press freedom” under the Trump administration, including threats to weaponize the FCC against unfavorable coverage, the dismantling of international broadcasters like Voice of America and Radio Free Europe, the detention and deportation of a journalist, and rising physical assaults on reporters covering demonstrations. According to the U.S. Press Freedom Tracker, journalist arrests rose from 15 in 2023 to 49 in 2024. Roughly one-third of American newspapers have closed since 2005, and more than 8,000 journalists have been laid off since 2022.⁴⁶

1
National Constitution Center. Interpretation: The Free Speech Clause
2
FIRE. Unprotected Speech: A Synopsis
3
Justia. Supreme Court Cases by Topic: Free Speech
4
First Amendment Encyclopedia. First Amendment Timeline
5
American Library Association. Court Cases Relevant to Censorship
6
U.S. Courts. Facts and Case Summary: Counterman v. Colorado
7
Supreme Court of the United States. Counterman v. Colorado, 600 U.S. ___
8
SCOTUSblog. Moody v. NetChoice, LLC
9
First Amendment Encyclopedia. Moody v. NetChoice, LLC
10
Supreme Court of the United States. Murthy v. Missouri, 603 U.S. ___
11
SCOTUSblog. Murthy v. Missouri
12
Missouri Attorney General. Consent Decree, Missouri v. Biden
13
Office of Senator Eric Schmitt. Schmitt Celebrates Historic First Amendment Victory
14
Supreme Court of the United States. TikTok Inc. v. Garland, 604 U.S. ___
15
Harvard Law Review. TikTok Inc. v. Garland
16
Supreme Court of the United States. Free Speech Coalition v. Paxton, 604 U.S. ___
17
FIRE. FIRE Supreme Court
18
Electronic Frontier Foundation. Section 230 of the Communications Decency Act
19
Harvard Law Review. Section 230 as First Amendment Rule
20
Roll Call. Bipartisan Bill Targets Government Censorship Threats
21
Electronic Frontier Foundation. New Bill Takes Aim at Government Pressure to Silence Lawful Online Speech
22
FIRE. 2026 College Free Speech Rankings
23
FIRE. 2026 College Free Speech Rankings
24
Inside Higher Ed. War on Student Speech
25
Scholars at Risk. Campus Protection Act Report
26
Texas Tribune. Texas Universities Campus Protests
27
ICNL. US Protest Law Tracker
28
NYCLU. Mahmoud Khalil v. Donald J. Trump
29
ACLU. Khalil v. Trump
30
ACLU. Defend Your Right to Learn
31
ACLU. Banned Books Week 2025
32
ACLU. Speech at Risk in America’s Schools
33
Freedom Forum. Artificial Intelligence and the First Amendment
34
Georgia Superior Court (via Loeb & Loeb). Walters v. OpenAI LLC
35
White House. National Policy Framework for Artificial Intelligence: Legislative Recommendations
36
Senate Commerce Committee. Cruz, Klobuchar Mark One Year of Take It Down Act
37
NAAG. The Benefits and Potential Shortcomings of the Take It Down Act
38
White House. Restoring Freedom of Speech and Ending Federal Censorship
39
Carnegie Endowment for International Peace. Taking the Pulse: Are Western Democracies Failing Free Speech?
40
European Parliamentary Research Service. EU-US Approaches to Online Speech Regulation
41
European Commission. The Digital Services Act
42
CSIS. Does the EU’s Digital Services Act Violate Freedom of Speech?
43
OHCHR. Expert Seminar: Links Between Articles 19 and 20 of the ICCPR
44
ARTICLE 19. Links Between Articles 19 and 20 of the ICCPR
45
Reporters Without Borders. 2026 RSF Index: Press Freedom at 25-Year Low
46
Reporters Without Borders. RSF Country Profile: United States