The Amendment That Should Have Broken the Ban

Australia banned social media for under-16s. The UK did the same. So did Turkey and Indonesia. Canada is moving toward its own version. Every one of those countries chose the same instrument: a categorical prohibition on access, enforced through age-gating at the platform or device level. None of those countries applies the constitutional test the United States does — strict scrutiny, which requires lawmakers to pick the least speech-restrictive option available. That test, if it worked as designed, should have forced a different question: is there a way to protect children that doesn't ban speech outright? The United States is the one democracy where that question is constitutionally unavoidable. And yet, as of this week, the pattern is reproducing almost exactly. California's Assembly unanimously passed a bill prohibiting platforms with addictive feeds from offering accounts to anyone under 16 ^[1]. South Carolina banned infinite scrolling, autoplay, and targeted advertising for minors on large platforms ^[1]. Minnesota's law prohibits specific design features — infinite scrolling, autoplay, push notifications — on child accounts ^[2]. The House passed the Kids Internet and Digital Safety Act 267-117 on June 29, with bipartisan support ^[3]. The Senate Judiciary Committee voted 22-0 to advance a ban on AI companion chatbots for minors ^[4]. US senators held a memorial honoring 270 children who died from suicide, cyberbullying, and drug poisoning linked to online harms, while pushing to align American law with Australia, the UK, Turkey, and Indonesia ^[5]. The political engine that drove bans through every other Anglosphere parliament is now running at full throttle in Washington. The First Amendment is in the room. It just isn't doing what it should. Here is what strict scrutiny, in theory, demands. When a law restricts speech, the government must show it serves a compelling interest and is narrowly tailored — meaning it achieves that interest through the least restrictive means. If lawmakers could protect children by requiring platforms to redesign their recommendation algorithms rather than by banning minors outright, the ban should fail. The less-restrictive alternative should win. That is the argument the First Amendment was built to force. It is not the argument anyone is making. Instead, the constitutional challenge is being deployed to block all regulation, not to redirect it. NetChoice — the trade group representing Meta, TikTok, and Snapchat — is challenging every state youth social media law on First Amendment grounds. It challenged Ohio's parental consent law and said it "will continue to fight the ruling" ^[6]. It called Minnesota's law "constitutionally defective legislation that will be enjoined by federal courts before it can take effect" ^[2]. It is fighting the Texas app-store age law that the 5th Circuit just allowed to take effect ^[7]. The ACLU is opposing state bans on First Amendment grounds too — but it is not proposing the design obligation as the constitutionally compatible alternative ^[8]. The constitutional argument is running in one direction: deregulatory. It is being used to strike down regulation of any kind, not to channel it toward the less-restrictive instrument that strict scrutiny should favor. The reason becomes clear when you watch what Meta does on the other track. While NetChoice challenges the laws in court, Meta's own lobbyists are pushing Congress and state legislatures to adopt app-store age verification — the same ban instrument, shifted to the operating-system level. A Meta lobbyist said the company is "very proud to support an app store age verification" ^[9]. Meta has told governments directly not to disrupt algorithms and to use device-level age enforcement instead ^[10]. The company is also lobbying Congress to insert a liability shield into KOSA that would grant immunity from state-law liability for minor safety in exchange for dropping opposition to the bill — what the American Association of Justice says would "eliminate pending lawsuits and block parents and schools from holding tech companies accountable" ^[11]. The strategy is legible: accept the ban, fight the design obligation, and use the First Amendment as a deregulatory tool to block anything that falls in between. This is why the design obligation — the one mechanism everyone acknowledges would address the actual harm — keeps falling away. The House bill passed 267-117 without the duty of care provision that the Senate's Blackburn-Blumenthal wing demands ^[3][12]. Sen. Marsha Blackburn criticized the House version precisely because it lacks a duty of care ^[5]. Sen. Blumenthal called it "a blank check to Mark Zuckerberg" ^[12]. The duty of care — an affirmative legal obligation to design platforms safely for children — is the provision that neither chamber will hold together for. Massachusetts Gov. Maura Healey's bill would "ban addictive algorithms that target youth," but that is a prohibition on a specific design choice, not a mandate to redesign the recommendation algorithm ^[13]. Minnesota's law comes closest to embedding design regulation, but its mechanism is feature prohibition, not an affirmative redesign mandate ^[2]. Every US state law, every federal bill, and every enforcement action — including the Mississippi-Roblox settlement with its specific design modifications ^[14] — stops short of requiring the one thing that would force platforms to change how their algorithms actually work. The circuit split is now real, and it is the vehicle that could bring this to the Supreme Court. The 5th Circuit stayed a district court injunction and let Texas's app-store age law take effect ^[7]. The 6th Circuit upheld Ohio's parental consent law 2-1, with Judge Clay finding the First Amendment burden "marginal" while Judge Ritz dissented that the law "restricts the ideas to which children may be exposed" ^[6]. Judge Robert Pitman, whose injunction the 5th Circuit stayed, framed the constitutional problem with a comparison that captures why this test matters.

The Act is akin to a law that would require every bookstore to verify the age of every customer at the door and, for minors, require parental consent before the child or teen could enter and again when they try to purchase a book. — Robert Pitman

That analogy — a bookstore carding every customer at the door — is the kind of reasoning that could yet strike down the ban instrument at the Supreme Court. The constitutional test is genuinely in motion, not settled. But here is the consequence worth watching: even if SCOTUS strikes down these laws, the political response would likely be a narrower ban, not a design obligation — because no American coalition is forming around the affirmative algorithmic redesign mandate that every government acknowledges and none enacts. The First Amendment, in the hands of the industry lobby and the civil liberties groups alike, is operating as a veto on regulation, not as a compass pointing toward the less-restrictive alternative it was designed to surface. The one lawmaker across any jurisdiction who explicitly framed the design obligation as the constitutionally compatible alternative is not American. Philippine Sen. Loren Legarda said the approach requires "enforceable duties for platforms and a framework that remains consistent with constitutional freedoms" ^[8] — the redirective argument the First Amendment should have produced in the United States, made instead in a Manila legislature that has no First Amendment to invoke. The irony is structural: the country with the strongest constitutional mechanism for forcing the less-restrictive alternative is the one where nobody is using it for that purpose. The countries without it chose the ban. The country with it is choosing the ban too — just through a different legal door. The Meta multi-state AG lawsuit, which survived Meta's motion to dismiss and goes to trial August 18, will test whether a jury can find that algorithms were "intentionally designed to be addictive to children" ^[15]. But even a plaintiff victory there would produce monetary penalties, not a binding injunction to redesign — consistent with what every court has done so far. The UK, meanwhile, is proposing to force platforms to "prioritize content from trusted news providers" in their recommendation systems ^[16] — an affirmative algorithmic redesign mandate, but for disinformation control, a domain where the government itself has skin in the game. The design obligation is politically viable when the imposing actor benefits from controlling the narrative. It falls away when the only beneficiary is a child. That gap — between what every government knows would work and what every government is willing to mandate — is the pattern. The First Amendment was supposed to be the exception that broke it. Instead, it is giving the same outcome a different legal dress.