We’re Not Ready
Bettina Krause September/October 2026In 2023, Supreme Court justice Elena Kagan sparked laughter in the courtroom during oral argument in a landmark tech case, Gonzalez v. Google. As counsel for both sides waded ever deeper into the complexities of social media algorithms, she stated the obvious: “We’re a court! We really don’t know about these things. We are not, like, the nine greatest experts on the internet.”
The law can’t keep up with communication technology and nowhere is this more evident than its messy intersection with free-speech norms and protections.
Consider, for instance, this odd AI fact. You will have more luck asking an AI chatbot to design a poster criticizing Donald Trump than if you ask for one criticizing China’s Xi Jinping. Meta’s Oversight Board released a study this year that found most AI models were more than twice as likely to refuse user requests to criticize repressive governments. And this spontaneous political censorship is happening to users in countries with strong free-speech protection.
What’s prompting this? Is it the data used to train the AI models? Could it be deliberate restrictions put in place by AI developers? Is it the indirect result of pressure from some governments? A combination of all these things? No one yet knows for sure.
We do know that whether we’re Googling for a quick answer, or having a back-and-forth with ChatGPT, we’re receiving vast quantities of information via platforms developed by private companies; platforms designed to operate globally, across jurisdictions with wildly different free-speech regimes.
From Hammer to Scalpel The challenge for the American legal system is that our constitutional framers failed to anticipate Google, Meta, Anthropic, or OpenAI. They drafted the First Amendment, prohibiting laws that “abridge the freedom of speech,” on the assumption that protecting free speech simply meant restraining the state from censoring its citizens. And, for centuries, this approach has made perfect sense. From the time that Johannes Gutenberg discovered how to make ink stick to metal letter blocks, public officials have regularly wielded the hammer of law to seize books and pamphlets, shut down presses, imprison printers and authors, and discourage disfavored religious or political viewpoints with hefty criminal penalties.
But today, a handful of private multinational corporations rival a state’s power to censor. In our new global public forum, these tech companies wield, not a hammer, but a scalpel as they silently rank, suppress, frame, moderate, and generate the information that lands in our feeds, or our interactions with AI chat bots.
And in responding to this, American lawmakers face a Catch-22.
If they under-regulate tech companies, a colossal amount of power to define “acceptable speech” is given to a few politically unaccountable, for-profit corporations. But if they over-regulate, then the harm envisioned by our First Amendment becomes a reality. The state itself imposes its own definitions of “acceptable speech” on private tech companies—and, by extension, on those who use their platforms.
America’s Legal Answer —So Far
In dealing with unprecedented legal questions posed by communication technology, the U.S. Supreme Court has so far doubled down on America’s distinctively strong free-speech protections. Two years ago, in the so-called NetChoice cases, the Court considered whether efforts by Texas and Florida to interfere with social media algorithms were constitutional. The Court said “No.” It suggested that when algorithms decide what appears in our social media feeds—choosing what content to suppress and what to promote—social media companies are engaged in constitutionally protected editorial discretion. It’s the First Amendment equivalent of a newspaper editor deciding what headlines will appear on the next day’s front page.
In Europe, however, lawmakers and courts are less squeamish about regulating tech giants and are heading down a different path. In 2022, the European Union adopted the Digital Services Act, which requires tech companies to design systems that comply with European definitions of harmful and illegal speech.
This means, to operate within the EU, tech companies must build platforms that can respond to the various speech laws of all 27 member nations. And each of these states have laws criminalizing some categories of speech—laws that would never pass constitutional muster in the United States.
Denmark, for instance, recently criminalized the “inappropriate treatment” of major religious texts—a response to incidents where the Quran was publicly burned. Or there’s Paragraph 188 of Germany’s criminal code, which provides for up to three years in prison for “insulting political figures.” And all EU nations, unlike the United States, have hate speech laws that criminalize the expression of some ideas or words deemed intolerant or offensive.
The Brussels Effect
It’s not only America and Europe that have taken distinctly different paths in defining and protecting free speech. Other jurisdictions—China, Iran, Russia, India, Saudi Arabia, Pakistan—implement far tighter controls over both speech and communication technology. Could these varied legal environments, and the need for tech companies to comply with local laws, be impacting how communication platforms function and interact with users in America? It’s possible.
Back in 2012, legal scholar Anu Bradford described a phenomenon called the “Brussels Effect.” She argued that European Union lawmakers in Brussels shape not just European, but global commercial practices. In broad terms, here’s how it works. First, multinational companies who want to trade in lucrative European markets voluntarily comply with EU regulations. Then, they realize that it’s cheaper and more efficient to simply adopt these strict EU standards across their entire global operations. And thus, you have the “Brussels Effect”—where lawmakers in one region are, in practice, designing de facto global standards for everyone.
It’s too early to say how far some form of the Brussels Effect may be operating within the world of communication technology, with tech giants engineering their global systems to respond to more restrictive speech laws. But it’s clear, in today’s interconnected digital environment, that America can no longer hermetically seal its borders when it comes to how it protects free speech.
A New Religious Liberty Frontier
We’re not ready for the communication revolution that’s already here. The tension we face is this: our First Amendment model isn’t fully equipped for a globally connected digital world. It’s not designed for private companies that create opaque algorithms and AI systems that can shape notions of acceptable speech or opinions on a vast scale.
On the other hand, tampering with our traditional First Amendment approach—which has created America’s uniquely strong free-speech regime—seems too foolhardy to contemplate.
Resolving this tension won’t be easy. For religious liberty advocates, it’s easy for our attention to be absorbed by familiar free exercise or establishment clause debates. But our portfolio has expanded. Free speech is the foundation on which our religious liberty rests. It rests on our ability to share our faith; to discuss ideas that others may want to punish as blasphemy, apostasy, or even just offensive; to dissent on religious grounds from majority opinions on everything from human sexuality to the relationship between religion and state.
Like Justice Kagan, we may not be “internet experts,” but it doesn’t take one to see that the challenge of defending our freedom of speech--and religion--online has only just begun.
Article Author: Bettina Krause
Bettina Krause is the editor of Liberty magazine.