What Actually Happened on August 5
The Senate Commerce Committee's August 5, 2026 markup produced a split verdict, not a clean sweep. Four bills cleared committee: the Kids Online Safety Act (S. 1748), the Youth AI Privacy Act (S. 4199), the CHATBOT Act (S. 4407), and the Children's Artificial Intelligence Toy Safety Act (S. 5171) — the last three by voice vote, per the committee's own release. The SCREEN Act (S. 737), by contrast, won a 15-13 majority but did not advance, because too few senators were present to satisfy the Senate's quorum rule — a procedural gap, not a substantive defeat, as EFF has noted. Chairman Ted Cruz framed the day's work as giving "parents more control and assurance that their vulnerable children won't be exposed to the worst possible elements of social media and AI chatbots"; Ranking Member Maria Cantwell said KOSA gives parents "the tools that help make them accountable" when platforms prioritize engagement over safety.
Why the SCREEN Act Is a Different Animal
It's worth being precise about what almost passed. The other three bills regulate specific product features — chatbot memory retention, parental account controls, design defaults for minors. The SCREEN Act would require any commercial site hosting even a fraction of adult content to verify every visitor's real-world identity, and it singles out traffic from known VPN addresses for additional scrutiny unless the operator can affirmatively show the user is outside the US. That is a fundamentally different mechanism: it doesn't ask platforms to change a feature, it asks them to build identity-verification and retention infrastructure for the entire adult internet, then extends extra suspicion to the exact tool millions of Americans use for ordinary privacy and security reasons — corporate VPNs, censorship circumvention, public Wi-Fi protection.
The Strongest Case for the Bill
Supporters aren't wrong that the status quo is genuinely bad. Self-attestation age gates — click a button confirming you're 18 — verify nothing and every parent knows it. A wave of state age-verification laws exists because legislatures concluded, reasonably, that a truly costless, unenforced gate is functionally no gate. And the SCREEN Act didn't invent the underlying legal theory from nothing: the Supreme Court's June 27, 2025 ruling in Free Speech Coalition v. Paxton upheld Texas's HB 1181, a similar age-verification mandate for adult sites, by a 6-3 vote applying intermediate rather than strict scrutiny — a genuine doctrinal shift that opened room for roughly two dozen states to pass comparable laws. A federal floor, proponents can argue, at least standardizes a patchwork rather than leaving each state to write its own version.
Where the Bill Overreaches
But Paxton is narrower than the SCREEN Act's proponents treat it as being. The Court's reasoning applies specifically to sites where sexual content is a substantial share of the offering — not to the general-purpose internet, and not to a rule that flags privacy-tool usage itself as suspect. The VPN provision is the tell: it doesn't verify age, it discourages the specific behavior — using a VPN — that privacy-conscious adults engage in precisely to avoid handing a service a persistent, breachable record of what they viewed and when. A bartender checks an ID and hands it back; the SCREEN Act's model is a database that keeps the ID, with no age-verification vendor guaranteed immune from being subpoenaed, hacked, or sold.
KOSA's Quieter Version of the Same Problem
The three bills that did advance aren't free of this tension either — they're just less blunt about it. KOSA's duty-of-care standard, as critics including EFF have pointed out, only functions if platforms can identify which users are minors, which pushes the same age-verification pressure through the back door rather than the front. And a liability standard tied to vaguely defined harms creates an incentive to over-remove lawful speech — including, ironically, the support communities and recovery forums that vulnerable teens rely on — rather than to make careful, defensible judgment calls about content.
The Proportionate Path
The honest reform agenda here is narrower than either side's current bill text. Congress should target verifiable design harms — engagement-optimized notification design, dark patterns, opaque data retention for minors' accounts — where the CHATBOT Act and Youth AI Privacy Act at least attempt something concrete, without conditioning access to lawful adult speech on a government-adjacent identity check. The SCREEN Act's quorum failure buys time, not a resolution; a revote is still expected. Before it returns, the bill needs two changes: drop the VPN-targeting provision entirely, and add real data-minimization and breach-liability teeth to whatever verification requirement survives. A law that makes every adult site keeper into an unwilling identity broker in the name of child safety solves one problem by creating a bigger one.