A rare bipartisan supermajority, a narrow path forward
On June 30, 2026, the House passed the Kids Internet and Digital Safety Act — the KIDS Act, H.R. 7757 — by a 267-117 vote, clearing the two-thirds threshold required under the expedited suspension process it was brought up on. The bill was then referred to the Senate Commerce, Science, and Transportation Committee on July 13, 2026, where a broader companion effort, the Kids Online Safety Act (S. 1748, sponsored by Sens. Richard Blumenthal and Marsha Blackburn), has been pending since May 2025. That the House found 267 votes for anything touching platform regulation is notable. That the bill now stalls in the Senate anyway is the more important story.
What the bill actually does
H.R. 7757 bundles several distinct interventions: a revised version of KOSA requiring platforms to adopt "reasonable policies" addressing harms like sexual exploitation, drug sales, and self-harm content aimed at minors; mandatory age verification for pornography sites; a requirement that AI chatbots disclose they are not human when interacting with minors; new restrictions on data brokers handling children's information; a ban on targeted advertising to minors; and limits on geolocation sharing and contact from unknown adults. Critically, the House version does not preempt state laws — states remain free to pass stricter rules — and it explicitly does not include language preempting state AI regulations, a provision most Republicans wanted and most Democrats rejected.
The strongest case for the Senate's version
Blumenthal and Blackburn's objection deserves to be taken seriously rather than dismissed as legislative posturing. Their Senate bill includes a duty-of-care standard — an affirmative, enforceable obligation for platforms to exercise reasonable care in product design to prevent foreseeable harm to minors, rather than merely publishing a policy. Their argument is that a policy requirement without a design-level duty lets platforms satisfy the letter of the law with a terms-of-service page while leaving addictive engagement mechanics, autoplay, and algorithmic amplification untouched. They called the House version's omission a choice that "will only maintain the status quo in which Big Tech uses our kids as profit centers." That is a fair description of what a policy-disclosure regime, unaccompanied by design obligations, can become in practice: compliance theater. If the goal is actually changing how platforms are built, not just how they're worded, a duty-of-care hook has real teeth that a reasonable-policies requirement does not.
Where the House bill still gets the trade-off wrong
Even granting that critique, the House bill's bigger problem isn't the provision it dropped — it's the one it kept. The mandatory age-verification requirement for adult sites lands on more solid constitutional ground than it would have a year ago: the Supreme Court's 2025 ruling in Free Speech Coalition v. Paxton, a 6-3 decision upholding Texas's age-verification statute, held that "no person — adult or child — has a First Amendment right to access speech that is obscene to minors without first submitting proof of age." That precedent makes a federal age-verification mandate for pornography sites far more litigation-resistant than it would have been under the stricter scrutiny of Reno v. ACLU. But constitutional does not mean sound policy. As the Electronic Frontier Foundation has argued, "every system demands users hand over sensitive personal information that links their offline identity to their online activity," and age-verification vendors have already suffered real breaches. A federal mandate multiplies that exposure across every adult site operating in the US, while doing nothing to address the design features — infinite scroll, algorithmic feeds, engagement-optimized notifications — that actually drive the compulsive-use harms lawmakers say they're targeting.
The better fix isn't either bill as written
The AI chatbot disclosure requirement is the part of this package that should be uncontroversial and easy to keep: telling a minor they're talking to software, not a person, is low-cost, informational, and directly responsive to a documented harm without touching speech content. The data-broker and targeted-advertising restrictions on minors' data are similarly narrow and defensible privacy measures that don't require guessing every visitor's age. Where both bills fail is on preemption. Refusing to preempt state law, as the House version does, guarantees platforms face a fifty-state patchwork of divergent age-verification, design, and disclosure rules — a compliance burden that entrenches incumbents with legal departments and locks out smaller competitors, the opposite of what child-safety advocates and innovation advocates should both want. A federal standard that combines the House's narrower, disclosure-based provisions with real preemption — while leaving duty-of-care liability for a separate, more carefully drafted bill that doesn't hinge on subjective "reasonable care" litigation — would do more for both kids and competition than either chamber's current draft. Given the Commerce Committee is now also juggling broader AI legislation in the same markup window, per reporting this week, a clean, narrow bill is more achievable than either side's maximalist version — but neither chamber has offered one yet.