The First Amendment at the Crossroads
The speech protections Americans take for granted are facing challenges from both left and right that the Supreme Court has not yet fully addressed. What happens next matters enormously.
The First Amendment is the most litigated provision of the American Constitution. It has also, in the past decade, become one of the most contested — not in the sense of being under political attack (though that is also true) but in the sense that the doctrinal framework courts have used to apply it is encountering new facts for which that framework was not designed.
The original architecture of First Amendment doctrine was built around a relatively simple model: the government seeks to suppress speech; the courts decide whether the suppression is constitutionally permissible. The suppression cases — prosecutions of war dissenters, restrictions on political organizing, censorship of sexually explicit material, punishment of flag burning — gave the doctrine its shape.
The cases now working their way through federal courts involve something more complicated: not the government directly suppressing speech but the government pressuring private companies to suppress speech; not individuals being silenced by state power but state officials being criticized for the power they exercise over speech; not censorship by fiat but censorship by platform architecture, by algorithmic curation, by the private exercise of power over digital infrastructure that determines what speech is visible to whom.
These are hard cases, and the courts' handling of them over the next decade will determine what free speech means in practice in the twenty-first century.
The jawboning problem
The First Amendment prohibits government from restricting speech. It does not, by its terms, prohibit government from asking private parties to restrict speech on its behalf — a practice legal scholars call "jawboning."
In the aftermath of the 2020 election and the COVID-19 pandemic, the Biden administration engaged in extensive communications with social media platforms regarding content they argued was harmful — election misinformation, vaccine misinformation, and related content. The extent and nature of these communications became the subject of Murthy v. Missouri (2024), in which the Supreme Court considered whether the government's communications with platforms constituted unconstitutional coercion of private actors.
The Court's decision in Murthy was a 6-3 ruling that held the plaintiffs lacked standing to bring the challenge — meaning that the Court did not reach the underlying constitutional question of when government communications with private platforms constitute First Amendment violations. The practical effect was to leave the doctrine in an unresolved state that permits continued uncertainty about where the line between permissible government communication and impermissible coercion lies.
Metaculus forecasts a 67 percent probability that the Supreme Court will issue a definitive ruling on the jawboning doctrine — establishing clear standards for when government communications with private platforms constitute coercion — before 2030. The case is waiting to be made. The question is which facts will provide the vehicle.
The platform as government actor
The mirror image of the jawboning problem is the government actor question: when do private platforms exercise so much power over public discourse that they should be treated, for constitutional purposes, more like governments than like private parties?
The historical doctrine — the "state action" doctrine — holds that the Constitution applies to government actors, not private ones. A government cannot prohibit flag burning; a private employer can prohibit employees from burning flags on company property. The doctrine made considerable sense when the primary venue for public discourse was public streets and government-licensed broadcast spectrum.
It makes less sense when two companies — Meta and Alphabet — control the primary infrastructure through which most Americans encounter political speech, when those companies make discretionary decisions about what speech is visible and what is suppressed, and when the suppression decisions have effects on political discourse that are comparable in scale to any government censorship regime in American history.
The Supreme Court's decisions in NetChoice v. Paxton and Moody v. NetChoice (2024), arising from Texas and Florida laws that attempted to prohibit platforms from moderating political speech, produced a fractured set of opinions that resolved the immediate cases without providing clear guidance on the underlying question of whether and how platforms can be required to carry speech they would otherwise moderate.
Justice Alito's dissent — arguing that at sufficient scale, private content moderation becomes the exercise of governmental power and should be subject to constitutional constraints — has not commanded a majority. But the argument it represents is not going away as platform concentration continues.
The AI speech problem
The most genuinely novel First Amendment challenge — novel in the sense that no existing doctrinal framework maps cleanly onto it — concerns the speech of artificial intelligence systems.
When an AI system recommends content, generates text, or declines to respond to a query, is that speech for First Amendment purposes? Does the First Amendment protect the AI's output? Does it protect the user's interest in receiving that output? When a government requires an AI company to include certain disclosures, or prohibits an AI system from discussing certain topics, or requires filtering of certain content types, what standard applies?
The existing framework is designed for human speakers and for the transmission of human-generated content. AI-generated content that has no human author — or that has a human prompt author and an algorithmic elaboration — fits this framework imperfectly. Courts have begun grappling with these questions; no consensus doctrine has emerged.
Kalshi was trading a contract on whether the Supreme Court will hear a case specifically involving AI-generated speech and the First Amendment before the end of 2028 at 54 percent. The probability seems right. The questions are arising faster than the doctrine is developing.
What the right and left both get wrong
The contemporary debate about free speech features two competing failure modes.
The conservative failure mode is to treat the First Amendment as a weapon against institutions that conservatives dislike — universities, media companies, social justice causes — while ignoring the ways in which the conservative political coalition has used government power to suppress speech it dislikes. Florida's "Don't Say Gay" law, Texas's social media regulation, the use of defamation litigation against critical journalists, the prosecutions of political opponents for speech that would clearly be protected in any administration whose political preferences were different — these are all speech restrictions. Their proponents do not describe them as speech restrictions; that does not make them less so.
The progressive failure mode is to treat harmful speech — false speech, hate speech, harassment — as categorically undeserving of constitutional protection, without adequately attending to the question of who decides what speech falls into the unprotected category and what happens when those decision-makers are not progressives.
The First Amendment's value is most evident not in cases where the protected speech is attractive or sympathetic but in cases where it is repugnant. The principle that government cannot suppress speech it dislikes is the principle that allows opposition to government policy, that protects political minorities, that makes accountability journalism possible. That principle does not become less important when the speech in question is false, offensive, or politically problematic. It becomes more important, because those are precisely the cases in which the political incentive to suppress is strongest.
Indicators to watch
— Free Speech Coalition v. Paxton: the Supreme Court's pending ruling on Texas's age-verification law for pornographic websites, which will clarify whether content restrictions applied to entire categories of protected speech receive strict scrutiny — The jawboning cases in the Fifth Circuit: new standing arguments on government communication with platforms — State AI speech legislation: a dozen states have pending legislation on AI disclosure, AI-generated political advertising, and AI content moderation that will generate the circuit splits that push Supreme Court review
Nathaniel Brooks is a contributing writer at The Auguro covering constitutional law, the Supreme Court, and the legal dimensions of technology and democracy.