Every country that regulates speech eventually discovers the same uncomfortable truth: laws written to shield people from hatred turn out to be remarkably well-suited for shielding governments from criticism. Even when legislation is drafted in good faith, its enforcement almost never is. Armenia is at the precipice of learning this lesson.

The Ministry of Justice has advanced a package that would extend liability for hate speech beyond the Criminal Code into civil and administrative law by creating new fines, fast-takedown duties, and conduct rules for the media. On top of that, it defines hate speech as anything that “humiliates, mocks, labels, or targets.” Not only are these concepts vague, leaving citizens unsure of what is off-limits, a law that turns on whether a sentence wounds, offends, or humiliates gives authorities broad discretion to punish speech they simply don’t like. That is the classic problem with vague speech laws. People who are unsure about what is illegal simply stop talking, and officials who are given sweeping powers tend to use them.

Crucially, “political opinion” is listed among the protected categories. In other words, the law would ban hate speech that “humiliates” or “mocks” someone’s political opinions or “worldviews.” But while hate speech laws are ordinarily justified as protection for historically vulnerable groups against dehumanizing attacks, they tend to punish political critics. Including denigration of “political opinion” — the one kind of speech that European law most emphatically protects — in hate speech legislation hands the government a tool to silence denigration of itself. It could also conceivably punish satire, which, by definition, mocks political opinions and worldviews.

And finally, the law would change how enforcement is designed to work. The draft imposes fines on media outlets and audiovisual service providers that fail to remove prohibited content, including content posted by users, immediately, but no later than three calendar days. Civil liability is expressly independent of criminal or administrative liability, meaning the same sentence can trigger all three at once.

No regulation should be read in a vacuum, and this one has arrived at a particularly tumultuous moment for the country. Over the past year, the Armenian government has conducted a visible campaign against its loudest institutional critic, the Armenian Apostolic Church.

The government of Armenian Prime Minister Nikol Pashinyan has moved repeatedly against its critics. Senior clergy of the Armenian Apostolic Church have been detained and arrested. Archbishop Bagrat Galstanyan, who led protests against territorial concessions, was detained on coup allegations; Archbishop Mikael Ajapahyan was sentenced to prison on similar grounds. The head of the Church, Garegin II, has faced travel restrictions and public calls for removal following disagreements with the government. The principle of separation of church and state is often described as mutual restraint. The state does not govern the church, and the church does not govern the state.

Armenia is a party to the European Convention on Human Rights, and Article 10 of that Convention is not a mere suggestion. Its core holding, repeated by the European Court of Human Rights for half a century, is that freedom of expression protects not only ideas that are welcome or harmless, but also those that offend, shock, or disturb. That directly conflicts with a statute whose operative verbs are “mock” and “denigrate.” Offense is not a recognized harm under Article 10; it is the price of living in a democracy, and Armenia must afford the same.

Calls for a Prime Minister to resign, criticism of territorial concessions, participation in protest movements, sermons that challenge the government’s handling of a war — these are not fringe cases of protected speech. They are the cases that legal doctrine is built to protect. They are, if the pun may be forgiven, canonical.

Armenia’s post-2018 “Velvet Revolution” was built on public demands for accountability, pluralism, and governance under law. Those commitments are written into the treaties Armenia has ratified, the Council of Europe institutions to which it belongs, and the constitutional text produced in the aftermath of that movement.

History has shown that, time and time again, when we sacrifice these commitments, freedom suffers and power consolidates. The question raised by this draft legislation is whether the legal order that emerged from a movement for pluralism is now setting up to police the boundaries of acceptable dissent.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

Subscribe now

Share

On Saturday, Federal Communications Commission Chairman Brendan Carr posted a warning to broadcasters who were running what he called “hoaxes and news distortions” about the Iran war, arguing they should correct course before their license renewals come up. “The law is clear. Broadcasters must operate in the public interest, and they will lose their licenses if they do not,” he added.

There’s just one small issue with this: the First Amendment does not have a fake news exception, and it never has.

The government cannot threaten to strip a broadcaster of its license because those in power dislike how a story was framed, what words a reporter chose, or whether a headline matched the administration’s preferred account of events. Even false speech is protected under the First Amendment, as the Supreme Court has repeatedly confirmed. The cure for journalism one finds “bad” is more speech, not a government official deciding whose reporting is fit to air.

Invoking the “public interest” standard does not cure the chairman’s First Amendment problem. The Supreme Court has been explicit that the public interest standard in broadcast regulation must be understood through the lens of the First Amendment, which must inform and give shape to how Congress and the FCC exercise regulatory power over broadcasting. Retaliating against news outlets because the President dislikes their coverage of an ongoing war is not a public interest rationale.

What if in 2003 the FCC threatened broadcasters’ licenses for unfavorable coverage of President Bush’s justification for the war in Iraq? I don’t have a time machine, but I do know that the outrage would be bipartisan. In 2020, the organization Free Press filed a petition to fine stations for alleged COVID misinformation under the related broadcast hoax rule. In a response, the General Counsel of the FCC rejected the request, explaining “...the Commission does not—and cannot and will not—act as a self- appointed, free-roving arbiter of truth in journalism.”

Carr has pointed to the agency’s longstanding news-distortion policy as the basis for his authority to make these threats. That policy is itself on constitutionally very shaky grounds and has been used rarely in the past 60 years. The FCC enforced it in just eight cases between 1969 and 2019, and those cases typically involved fabrications of news events, not editorial judgments about what to emphasize or how to describe a situation. Carr himself said in 2019, “Should the government censor speech it doesn’t like? Of course not. The FCC does not have a roving mandate to police speech in the name of the ‘public interest.’”

It’s clear that the Chairman is aware of the First Amendment’s obstacle. But statements like the one from last week are not there for legal follow-through; they are there for the chilling effect.

When a government official tells broadcast stations to correct their coverage or face consequences, owners of those stations have to decide if they want to continue reporting in a way that will please the government or take the very expensive legal and reputational gamble of going against the government to maintain their ability to operate. It’s important to remember that the government does not need to revoke a single license to change what gets reported, how a story gets framed, or whether a producer decides a particular piece is worth the headache. They just need to flex their muscles and point the editorial winds in the direction they want them to blow.

That is why the Supreme Court, case after case, has held that government pressure designed to alter private speech raises the same constitutional concerns as outright censorship. Most recently, in National Rifle Association v. Vullo, the Court ruled, “Government officials cannot attempt to coerce private parties in order to punish or suppress views that the government disfavors.” A regulatory chairman posting public warnings to news organizations the day after the president complains about their war coverage fits that description with uncomfortable precision.

Carr has suggested that broadcasters who do not like these conditions are free to move to streaming, which the FCC does not regulate. That is a remarkable concession since he is acknowledging that the government can use the broadcast licensing process to extract editorial compliance that it could never demand of any other publisher. The scarcity rationale (limited airwaves that require government-granted licenses and allocation) that once justified greater government involvement in broadcast content has been eroding in the courts for decades and can no longer bear the weight being placed on it here.

The First Amendment does not exist to protect popular speech or coverage that aligns with government narratives. It was written precisely for moments like this one, when officials with power over the press are unhappy with what the press is reporting.

The broadcasters covering this war are doing their jobs. Is Chairman Carr doing his?

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

In the hours after a tragedy, the public response now follows a familiar and horrifying cadence: breaking news, statements from officials, and public mourning. And very quickly, social media’s attention shifts from what happened to who has spoken and who has not. “Your silence is deafening” becomes a common refrain.

Followers scroll through stories not only for information, but for evidence of moral alignment. When a creator fails to post about a tragedy, some interpret the absence as indifference, complicity, or cowardice.

This expectation has hardened into a social mandate. And it raises an uncomfortable question about our speech culture — when did speaking become an obligation rather than a choice?

Recently, the killing of Alex Pretti prompted this exact cycle. Followers immediately demanded statements from influencers whose platforms had nothing to do with politics or public policy. When some did not post, backlash followed. The message from the online audience was clear: a celebrity or influencer’s silence can only be interpreted as very loud speech in favor of a view opposite to mine.

After the 2024 U.S. election, fitness influencers found themselves in precisely this bind. As Men’s Health reported, creators whose content focused entirely on workout routines, nutrition advice, and physique updates were suddenly facing demands to declare their political positions. Some complied, posting carefully worded statements. Others stayed silent and watched their comment sections fill with accusations. One influencer told the magazine that followers were combing through his old posts looking for evidence of his politics, analyzing which brands he promoted, and scrutinizing who he followed. In this polarized digital landscape, you must be a political actor, whether you want to be or not.

This represents a fundamental shift in what audiences believe they are owed. A decade ago, followers might have been satisfied with workout tips and motivational content. Now, they want ideological transparency. They want to know not just how their favorite creator trains, but how they vote.

This is where the conversation moves beyond influencer drama and into genuine questions about what free expression means in our culture. In the United States, free speech is not just the right to speak. It is also the right not to. The freedom to choose when, how, and whether to comment is foundational to expressive liberty.

Free speech doctrine, particularly in the American tradition, views compelled speech as one of the most dangerous forms of government overreach. But what happens when compulsion comes not from the state but from the crowd — when the punishment is not legal but social ostracism or economic loss?

Influencers are private individuals with public reach; they are brand-builders rather than elected officials. Yet they are increasingly treated as if they owe their followers civic leadership. Their audiences fund their visibility, and in return, many followers feel entitled not only to content but also to conscience. Followers view their favorite influencers’ platforms as megaphones to amplify their own grievances and causes.

This is where parasocial relationships become philosophically significant. The term “parasocial” was coined by sociologists Donald Horton and Richard Wohl in 1956 to describe the one-sided intimacy audiences develop with public figures. The follower knows the influencer; the influencer does not know the follower. Yet the relationship feels mutual.

Subscribe now

The influencer shares vulnerable moments, speaks directly to the camera as if to a friend, solicits opinions, and responds to comments. The follower invests emotionally and often financially. The expectation, then, is rooted in a category error: treating a parasocial relationship as if it were a communal one. Followers expect influencers to be part of their moral community, someone who shares their values, their outrage, and their grief. And in communities, silence during moral crises is noticed and judged.

Here’s the other side of it — influencers trade on authenticity. The most successful ones cultivate trust with their audience and build a personal bond. When tragedy strikes, followers want reassurance that the people they support — and whose lifestyle their engagement funds — also share their values. That impulse is understandable. But there’s a point at which expression stops being authentic and starts being performative. The demand for instant statements often produces checkbox activism.

The demand for speech, then, becomes a kind of loyalty test. It is not enough to refrain from saying harmful things; one must actively affirm the right things. Silence is no longer neutral. It is no longer private. It is no longer a legitimate exercise of negative liberty. Instead, it becomes evidence of hidden vice, proof of moral deficiency, and a gap in the performance that reveals the “real” person underneath.

Consider the case of Alix Earle, a TikTok creator often suspected by her audience of holding conservative political views. When she posted two Instagram stories about a Minnesota tragedy, the online reaction was a surprise that served as social proof, “Even Alix Earle posted.” The implication is that even if someone who is supposedly ideologically unaligned speaks out, everyone else’s failure to do so becomes more damning.

But often, these are cases of compelled speech by social coercion. And it poses a genuine threat to expressive freedom — not because any individual demand for a statement is totalitarian, but because the cumulative effect is a culture in which speech becomes obligatory, silence is punished, and the line between authentic expression and performative compliance disappears entirely.

None of this is an argument that influencers shouldn’t speak. Many do so thoughtfully and responsibly, using their platforms to amplify critical information, fundraise, and educate. But we should ask ourselves what we’re actually creating when we demand speech on command — and when we interpret silence as endorsement for a particular viewpoint. A culture that punishes silence doesn’t make us better informed or more united. It makes us performative. If someone you follow for their makeup tutorials doesn’t share your views on a political tragedy, would you rather know that through authentic silence or a PR-approved statement designed to offend no one?

Influencers typically have expertise in specific niches. There’s no reason to expect that they would have competence in the nuances of politics or policy. Their first encounter with an issue often occurs at inflection points when their audience demands an opinion, a statement, or an action. When people are forced to perform solidarity on an issue they are not well informed about, the pressure may lead them to reject that position, and they might be pushed in the opposite direction.

In The Coddling of the American Mind, Greg Lukianoff and Jonathan Haidt argue that public shaming and call-out campaigns tend to increase polarization rather than reduce it and can push people towards defensive tribalism. Research on prejudice reduction (from Gordon Allport’s 1954 The Nature of Prejudice to modern studies) shows that continuous meaningful contact and dialogue produce durable attitude change. But in a world of shortened attention spans and instant-gratification expectations, that’s not a solution used by most.

Free speech was never meant to guarantee that everyone speaks on cue. It was intended to ensure that, when we do, it is of our own volition. And if we, as listeners or viewers, don’t like what someone chooses to say or not say, we have our answer about whether we want to continue following them. That’s the marketplace of ideas working.

The question isn’t whether influencers have the right to stay silent. They do. And sometimes silence can be just that: silence. The real question is whether we are building a culture where that right can actually be exercised, or have we created a system where silence is so heavily punished that speech becomes compulsory in every circumstance?

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

Last week, the National Football League fined Houston Texans linebacker Azeez Al-Shaair nearly $12,000 for wearing eye black that read “Stop Genocide” during a playoff game against the Pittsburgh Steelers. The league said the message violated its uniform policy, which bars “personal messages” on game-day equipment.

But just months earlier, in the same season, the NFL permitted tributes to Charlie Kirk, the conservative activist and Turning Point USA founder, across the league after his assassination. The league chose to hold a moment of “silent reflection” during Thursday Night Football and authorized the teams to hold their own tributes if they chose to.

These recent events illustrate how the NFL has never been fully comfortable with players and teams sharing opinions, while at the same time adopting policies that seem to take overt political stances. And it has nothing to do with whether one agrees with Al-Shaair’s characterization of the events in Gaza or supports Kirk’s views. Instead, it’s about a league with immense cultural power and influence, normalizing that selective censorship can be a virtue rather than a problem.

The NFL insists that it is neutral; its actions suggest otherwise.

Neutrality has become a talking point and a corporate marketing tool that creates a perverse incentive structure. Players learn quickly that some viewpoints carry little risk while others may cost them money, opportunities, or career longevity. This inconsistency in the implementation and enforcement of the league’s policies is especially glaring given the NFL’s global nature.

Defenders of the NFL may argue that “Stop Genocide” is inherently more controversial than honoring a political commentator. But that argument only exposes the core problem. Who decides what counts as inflammatory? Which figures are mainstream enough to merit public tribute? These judgments inevitably reflect the values and biases of those in power. When they are applied unevenly, they stop looking like rules and start looking like preferences.

Ten years ago, Colin Kaepernick, the quarterback for the San Francisco 49ers who took the team to the Super Bowl, knelt during the National Anthem to raise awareness of racial inequality and to protest police brutality. This sent shockwaves through the NFL and shaped conversations on free speech that year. After becoming a free agent, Kaepernick wasn’t able to sign with another team and sued the NFL, alleging that the league had blacklisted him. They later reached a settlement, and the NFL adopted a new national anthem policy requiring players to stand during the anthem or remain in the locker room.

It becomes even more difficult to square the NFL’s inconsistent approach to political “neutrality,” given its relationship with the U.S. government. In 2015, a year before the Kaepernick controversy, a PBS investigation revealed that the Department of Defense paid the league more than $5 million in taxpayer money between 2011 and 2014 to stage “patriotic” displays at games, including on-field ceremonies and in-stadium tributes to soldiers.

The NFL was comfortable accepting taxpayer funding to present a specific vision of patriotism and national identity. But when players engage in unscripted, bottom-up political expression, the league suddenly rediscovers an aversion to political messages.

Subscribe now

To be clear, the NFL is a private employer. It has the legal right to regulate on-the-job expression and promote whatever ideas it wants. But with that great power comes great responsibility, especially for an institution that occupies a significant cultural space in American life.

The league could establish clear, consistently enforced rules — either prohibit all political expression during games or create genuine space for players to express their convictions. Instead, it has chosen a third path, selective enforcement, that protects certain speech while punishing others, all while claiming to stand for nothing.

The National Football League, which markets itself as a reflection of America and its values, has chosen a path that mirrors viewpoint discrimination, the form of censorship the First Amendment prohibits in government, precisely because it corrupts the marketplace of ideas. An inconsistent posture on free speech doesn’t reflect those values.

Free speech is a legal guarantee in the United States, but it’s also a social and cultural value. We must look to our institutions, like the National Football League, to uphold and reinforce it.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

Imagine walking into the world’s largest library. You’re surrounded by millions of books, but there is no catalog, no librarian, no sections. Every book is randomly stacked, with no indication of subject or relevance. You’re left to wander and will probably never find what you’re looking for. That’s the Internet without algorithmic curation. And it’s the future we’re hurtling toward if legal protections like Section 230 and First Amendment editorial rights are stripped away.

Today’s Internet only works because algorithms help us find the book in the endless stacks. Platforms sort, filter, and recommend content based on user behavior and content policies. They help enforce safety rules at scale, personalize feeds so users find what matters to them, and create order in the chaos. Just as a wise librarian guides a teenager to age-appropriate materials, algorithms help steer users toward content that suits them. Our research associate Isabelle Anzabi has put together a short and helpful explainer on how these recommendation systems work.

Unfortunately, a growing chorus of voices is arguing that the legal foundations that built the Internet as we know it need to change. They contend that platforms are either moderating too much or too little content, and that the algorithms used to perform this function should be subject to government scrutiny. These arguments often stem from a misunderstanding about how our current legal framework has shaped the modern Internet and what would happen if we were to abandon it.

Courts have repeatedly affirmed that editorial rights apply to decisions about what not to say, just as much as what is said. Whether it’s a newspaper refusing to print something the government requires them to, a parade choosing who can be a part of it, or a utility company declining to include outside messages in its mailings, the principle is the same: private entities cannot be compelled to carry speech they do not endorse. This principle extends to social media platforms and their algorithmic systems. When a platform chooses to promote one post over another, or enforce community standards, it is making a choice about speech, and that choice is constitutionally protected.

Section 230 reinforces this by shielding platforms from liability for third-party content, whether they host, organize, or moderate it. Without this protection, platforms would face lawsuits every time they tried to enforce rules or highlight content. Before Section 230 existed, courts penalized platforms that attempted to moderate content. This framework created an incentive for platforms to moderate everything or do nothing at all, allowing content that would be potentially sensitive to anyone in the country. Section 230 reversed that, empowering platforms to act responsibly without fear that every decision could end in court.

But a recent decision from the Third Circuit threatens to unravel this framework. In TikTok v. Anderson, the court held that when an algorithm promotes harmful content, it becomes the platform’s own speech, thereby stripping away Section 230 immunity. If Section 230 protections or First Amendment editorial rights are rolled back, the consequences could be profound.

Some platforms will over-moderate to avoid legal exposure, removing lawful but controversial content. Others will under-moderate, allowing all types of content, including things most of us wouldn’t want to see, to spread unchecked. Such a shift will not harm the powerful but the vulnerable, the dissenters, and the voices that depend on intermediaries to be heard. Smaller platforms and start-ups may shut down or avoid hosting speech and change their business models altogether due to litigation risk.

In a new policy paper for The Future of Free Speech, I argue that algorithmic curation is not just a technical necessity; it is a form of editorial discretion protected by the First Amendment that is essential to the functioning of the digital ecosystem. Section 230, in turn, provides the statutory safety net that makes such moderation feasible without constant legal peril.

Read It Here

Without algorithms, we would be forever lost in the library stacks, unable to find and consume the content that matters most to us. If we care about free speech online, we should fight to preserve the legal protections that make algorithmic curation possible.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

When FCC Chairman Brendan Carr questioned ABC’s and local affiliates' licenses over Jimmy Kimmel’s monologue and Disney responded by pulling Kimmel’s show off the air, the headlines focused on outrage, hypocrisy, and political overreach.

But the most dangerous part of this episode isn't just that a government official attempted to silence a comedian. It's that everyone watching, from station executives to law students to future public servants, just learned that the First Amendment is not invincible.

This is not how American constitutional democracy is supposed to function. In NRA v. Vullo, decided just last year, the Supreme Court unanimously reaffirmed that government officials cannot use threats of enforcement to suppress speech.

Carr’s response was not simply criticism; it was coercion. On a podcast, he explicitly suggested that, unless Disney and its affiliates “took action” against Kimmel’s show, his agency would “do this the easy way or the hard way. These companies can find ways to change conduct and take action, frankly, on Kimmel, or there’s going to be additional work for the FCC ahead.” The Chairman also discussed several tools the FCC could utilize, including the public interest obligation and the “broadcast-hoax” and “news distortion” rules.

Subscribe now

That was a targeted, strategic use of regulatory power to threaten protected speech. And it worked. Within hours, Nexstar, the owner of a significant number of local ABC affiliates (which has a merger, valued at $6.2 billion, pending before the FCC), announced it would no longer air Kimmel’s show.

Shortly after, Disney suspended him “indefinitely.” This is censorship by proxy — a modernized version of the heckler’s veto, powered by regulatory intimidation rather than mob outrage. And the damage doesn’t stop with Kimmel.

Because somewhere right now, a college student is watching all of this unfold. A Hill staffer is taking notes. A law student is applying to judicial clerkships and sifting through old telecommunications laws, wondering where the precedent lands. And the lesson they are absorbing is this: When you’re in power, you can punish speech you don’t like. You just have to be creative about it.

The “public interest” standard can be manipulated. Licensing threats can be framed as oversight. And the public will eventually move on. In short, they are learning how to break the First Amendment from within.

The long-term danger here is generational, cultural, and institutional. If today’s young professionals grow up in a world where comedy gets censored by bureaucrats, where speech is policed with FCC leverage, and where the loudest defenders of “free speech” conveniently go silent when their political allies do the censoring, then the foundational norms that protect speech will erode. They will inherit not a robust marketplace of ideas, but a fragile, conditional system where speech survives only if it doesn’t offend the ones in power.

We cannot afford to treat this as another partisan fight. This is about what we are normalizing and what comes next. The FCC does not have the power to punish jokes that land poorly or speculation that ends up being wrong. Quite the opposite — the First Amendment gives “breathing space” protecting false statements and hyperbole that are “inevitable in free debate.”

Our constitutional system will withstand and protect the speech rights that are under attack today. But today’s students are tomorrow’s policymakers. If we want them to inherit a First Amendment worth defending, we can’t treat it like an afterthought today.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

At a time of deep partisan divides in the wake of last week’s horrific murder of Charlie Kirk, it was starting to seem like nothing would bring us together. That is, until a video of U.S. Attorney General Pam Bondi was posted online outlining her view of where lines should be drawn on First Amendment-protected speech.

On a podcast, Bondi said, “There’s free speech and then there’s hate speech,” and warned that her office would “go after” individuals engaging in the latter. When the clip was posted to social media, it was immediately ratioed by commentators from all sides, who rightly argued that hate speech is free speech under the First Amendment.

Regardless, it appears the administration is doubling down on its unconstitutional stance.

In a follow-up post, Bondi clarified that when hate speech “crosses the line into threats of violence,” it is not protected by the First Amendment and cited federal statutes that criminalize threats against individuals and public officials.

When asked about her comments, President Trump responded, “We’ll probably go after people like you because you treat me so unfairly,” referring to a reporter. “You have a lot of hate in your heart,” he added. “Maybe they’ll have to go after you.”

The idea that hate speech is unprotected has usually come from the other side of the political spectrum. But it’s critical, especially at a time like this, to understand that the instinct to punish speech because it feels hateful or unfair is precisely what the First Amendment was written to restrain. The Constitution does not require us to approve of hateful speech. It requires the government to stay out of the business of deciding what speech is allowed.

Subscribe now

The Supreme Court has been clear on this point. In Snyder v. Phelps, the Court upheld the rights of protesters to express vile views at a military funeral. In Matal v. Tam, the Court unanimously struck down a federal trademark law that discriminated against “disparaging” words. Justice Alito, writing for the Court, said that “the proudest boast of our free speech tradition is that we protect the freedom to express ‘the thought we hate’”.

Let’s talk about when speech is actually unprotected. As established in Brandenburg v Ohio, speech intended and likely to provoke imminent violence or law-breaking is unprotected. For example, urging a mob to attack a minority group (for example, if a mob of Neo Nazis is urged to attack a Jewish synagogue) right now would not be protected. But advocating hateful ideas in general, or talking about violence in some abstract future sense, is protected, because the incitement exception is limited to immediate calls for unlawful action. Cheering for someone’s death, or saying that that person deserved to die, is also constitutionally protected speech. Statements meant to seriously frighten or intimidate someone with violence are not.

The line is drawn between merely offensive rhetoric and genuine threats of harm. If a speaker directs hateful words toward an individual or group in a way that a reasonable person would take as a serious intent to do violence, that crosses into unprotected territory. For example, a racial slur is protected, but if a Mafia boss says, “I’ll break your legs if you don’t keep quiet,” that’s a true threat.

These distinctions are very narrow and don’t give the government power to decide what speech is good and what speech is bad. If government officials are allowed to decide what counts as hate speech, then speech becomes a political weapon they can use to enforce their ideological preferences on Americans. What happens when a future administration claims that strong religious beliefs or criticism of public health policies amount to hate? The First Amendment exists precisely to prevent this kind of viewpoint-based enforcement.

Conservatives have long defended free speech as a check on government power. That commitment should not falter when the rhetoric becomes uncomfortable. Because if the government gains the power to prosecute disfavored viewpoints, there won’t be a neutral way to draw the line. Today’s hate speech prosecution is tomorrow’s speech code against your values.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

California lawmakers are once again testing the constitutional limits of internet regulation, and SB 771 is their latest attempt to sidestep Section 230 and the First Amendment. On its face, the bill claims to impose liability on large social media companies only when they materially contribute to violations of longstanding civil rights laws, such as the Ralph and Bane Acts.

In reality, SB 771 seeks to punish platforms for the content they algorithmically recommend, even if that content is lawful, protected speech, by labeling the recommendation itself as “conduct” rather than “speech.” That semantic shift doesn’t cure the constitutional problem; it magnifies it.

SB 771 tries to argue that when a platform’s algorithm amplifies a third party’s violent or harassing post, the platform is no longer just a conduit but a participant. That theory flies in the face of Section 230’s core promise: that platforms are not liable for third-party content they didn’t create, even if they amplify it.

Section 230(c)(1) makes it crystal clear that “no provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another.” Courts across the country, including the Ninth Circuit, have interpreted this immunity broadly, and for good reason. Stripping it away just because the content appeared in a user’s feed thanks to a ranking algorithm ignores how the modern Internet works. Algorithms are not optional features but the structure of the digital world. To call that structure unlawful conduct every time someone posts something harmful is to demand omniscient pre-censorship on a massive scale.

The bill’s backers cite the Third Circuit’s recent decision in Anderson v. TikTok to argue that algorithmic curation can fall outside of Section 230. But that case isn’t binding in California. In fact, it clashes sharply with Ninth Circuit precedent, particularly Dyroff v. Ultimate Software and Gonzalez v. Google.

In Dyroff, the court held that algorithmic recommendations, even those that allegedly pushed users toward dangerous content, were protected under Section 230 because they were merely tools for publishing third-party speech. SB 771 essentially invites plaintiffs to relitigate that question under a thin veneer of civil rights enforcement. But this amounts to jurisdictional forum shopping disguised as state-level accountability.

Even if the bill managed to dodge Section 230 preemption, it still runs headfirst into the First Amendment. The Supreme Court has repeatedly affirmed that editorial discretion is itself a protected form of speech. As Moody v. NetChoice reaffirmed, platforms exercise expressive judgment when they choose what content to surface, bury, or remove. SB 771 punishes that judgment by tying massive civil penalties to algorithmic outputs that correlate with alleged statutory violations. But algorithms just operationalize editorial judgment at scale.

Subscribe now

The legislative analysis tries to reframe this as a modest effort to prevent truly egregious harms, violence, intimidation, and harassment. But the liability hook is still user speech. And the platform’s so-called misconduct is still choosing to display that speech in a feed. No matter how much the drafters insist this is about conduct, not content, the First Amendment doesn’t allow that kind of end-run. Calling curation “product design” doesn’t change the constitutional character of what’s being punished.

Worse, SB 771 would chill lawful speech far beyond the limited category of threats or harassment it targets. Platforms, faced with the risk of uncapped civil penalties for unknowingly recommending third-party posts that might be construed as violations of state civil rights laws, will overcorrect. We have seen this play out over and over again abroad. They’ll suppress content, delete controversial users, and flatten political and cultural discourse to a dull hum. The law doesn’t require them to, but the litigation and liability risk all but guarantee it.

This bill is not about safety. Instead, it represents another effort by a state government to circumvent platforms' constitutional protections in order to implement state-controlled content moderation. And it sets a dangerous precedent: that a state can impose crushing liability on private actors for failing to sufficiently sanitize their public forums (in a manner that a specific state wants them to), even when the offending speech is constitutionally protected.

That idea is un-American, unconstitutional, and technologically incoherent. It deserves to be rejected not just because it conflicts with Section 230, but because it betrays the First Amendment’s most basic promise: that the government cannot decide what speech is good and what speech is bad.

Share

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

In his March address to a joint session of Congress, President Donald Trump praised a bill that would go after the use of AI deepfakes to create “revenge porn,” adding "I’m going to use that bill for myself too, if you don’t mind, because nobody gets treated worse than I do online, nobody.”

Few would deny that the creation of nonconsensual intimate images and videos is one of the most worrying developments in our AI era. But when a president whose administration has repeatedly challenged and pushed the boundaries of the First Amendment makes such proclamations, even in jest, we should carefully consider the levers of power we are granting to the government to police such content.

On April 28, Congress overwhelmingly passed the bipartisan Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks (TAKE IT DOWN) Act. The legislation, now awaiting President Trump’s signature, criminalizes the nonconsensual distribution of intimate images (NDII), whether real or digitally altered, and requires a notice and take-down mechanism.

At its core, the TAKE IT DOWN Act is a well-intentioned effort to address nonconsensual distribution of intimate images. But beneath its protective promise lie serious threats to free expression and even the rights of the very individuals it aims to protect. The Act’s vague definitions, sweeping authority, and lack of meaningful safeguards open the door to unconstitutional censorship, while also falling short of combating the real harms of NDII.

A System Without Guardrails

The bill’s notice and removal mechanism requires companies to remove any content described as an “intimate visual depiction” within 48 hours of receiving a takedown request from either the person depicted or their authorized representative who claims it was published without consent. While the legislation’s criminal provisions carve out exceptions for consensually shared explicit images and for content used in medical, legal, or educational contexts, the notice and removal provisions do not.

This absence of nuance, coupled with the 48-hour constraint, encourages platforms to rely on automated filtering systems already prone to over-removal. The Act also does not require people to file requests under penalty of perjury, and it does not punish false claims. This means people can report content simply because it doesn’t align with their personal beliefs, even if it is not private or sexual at all. As a result, platforms are likely to take down speech even if it is consensual and constitutionally protected.

The Digital Millennium Copyright Act (DMCA) has a somewhat similar notice and takedown system, but unlike the TAKE IT DOWN system, it requires complainants to attest they are authorized to file, has a counter-notice system, and imposes liability on those who made knowing material misrepresentations. Even with those safeguards, the DMCA system has reportedly still been used to silence legal criticism and expression. The system under this new law will undoubtedly overwhelm tech companies, threatening the ability of smaller platforms to operate and hindering justice for NDII victims.

Equally troubling, this law also prohibits counterclaims against platforms that act “in good faith” on takedown requests, regardless of whether the content is lawful or not. The Cyber Civil Rights Initiative clearly explains how this removes any avenue of redress for individuals harmed by the removal of protected speech while eliminating any meaningful incentive for platforms to safeguard legal expression when responding to takedown demands.

A New Government Tool

The TAKE IT DOWN Act vests significant enforcement authority in the Federal Trade Commission (FTC), empowering the agency to treat noncompliance with the takedown provisions as an “unfair or deceptive act or practice” under Section 5 of the FTC Act. This grants the FTC broad discretion to investigate, penalize, and potentially litigate against online platforms, even in cases where takedown requests are legally or factually dubious. This is particularly concerning in an administration that has repeatedly weaponized legislation to target media outlets on ideological grounds.

Overbreadth and Ambiguity

The Act’s overbreadth exacerbates these concerns. While its criminal provisions adopt a relatively narrow and constitutionally sound definition of nonconsensual intimate imagery (NDII), the takedown provision applies to a far broader category: any “intimate visual depiction” alleged to be nonconsensual by an “identifiable individual.” The law offers no clear standard for evaluating such claims, no requirement for platforms to assess their validity, and no meaningful distinction between authentic images and those that are fabricated or manipulated.

Congressional efforts to protect victims of non-consensual intimate imagery are incredibly important and laudable. But good intentions do not always yield good policy and constitutional laws. Protecting victims and protecting free speech are not mutually exclusive, and we must be more vigilant now than ever to safeguard both.

Ashkhen Kazaryan is a Senior Legal Fellow at The Future of Free Speech, where she leads initiatives to protect free expression and shape policies that uphold the First Amendment in the digital age.

Ashley Haek is a communications coordinator at The Future of Free Speech.

Share