The House passed H.R. 7757, the KIDS Act, by a vote of 267-117 on June 29, 2026, folding the Kids Online Safety Act (KOSA), COPPA 2.0, the SAFE Bots Act, the Safer GAMING Act, the SPY Kids Act, and roughly a dozen other child-safety bills into a single package. Sponsored by Rep. Brett Guthrie (R-KY) and introduced March 3, 2026, the bill now moves to the Senate, where — as Tech Policy Press reported the day of passage — reconciling it with the Senate's own KOSA text faces "long odds" before the August recess and midterm elections.
The bill's text includes an explicit rule of construction stating that nothing in it shall expand, limit, or alter the meaning of Section 230 of the Communications Act of 1934. That single sentence has let sponsors claim the bill leaves intermediary liability alone. It is true as written. It is also close to beside the point.
The Case Advocates Are Making
The underlying harms driving this bill are real and well documented: platform-design features that reward compulsive use, algorithmic amplification that connects predatory adults to minors, sextortion schemes run through direct messages, and AI chatbots that misrepresent themselves to children in crisis. State attorneys general have spent years building records on these harms, and a bipartisan majority spanning 267 members — well past the two-thirds threshold needed to pass a bill under suspension of the rules — is a real signal that both parties see the status quo as untenable. Public Knowledge, a group that has criticized earlier KOSA drafts, endorsed this version specifically because it targets "protective defaults" like limiting recommendation algorithms and disabling geolocation for minors rather than regulating what content is legal to view.
Duty of Care, Removed — Then Rebuilt Piecemeal
The most consequential change from earlier KOSA drafts is the removal of the generalized "duty of care" that would have required platforms to mitigate broadly defined psychological harms. That provision survives in the Senate's version (S. 1748), which still includes mental-health-related harms the House measure dropped. House negotiators cut it after years of warnings — including from EFF — that a duty of care standard would push platforms to over-remove lawful speech to avoid open-ended liability.
In its place, the House bill requires platforms to "establish, implement, maintain, and enforce reasonable policies" addressing a specific, enumerated list of harms: sexual exploitation, substance use content, gambling, and financial fraud aimed at minors. That is narrower than a duty of care in name. But EFF's June 24 analysis flags the mechanism that recreates the same pressure: platforms face liability if they "knew or should have known" a user was a minor, a standard that leaves courts and regulators to decide after the fact whether a platform should have caught it. Facing that uncertainty, EFF argues, "many services will inevitably choose to remove that speech or restrict those discussions" rather than risk litigation — the same over-removal dynamic duty-of-care critics warned about, arrived at through a different legal doorway.
"Faced with legal risk, many services will inevitably choose to remove that speech or restrict those discussions." — EFF, June 24, 2026
A Section 230 Carve-Out in Substance, Not in Name
This is the real story behind the Section 230 rule of construction. Section 230 immunizes platforms from liability for hosting third-party content and for their editorial choices about that content. The KIDS Act doesn't touch that immunity on paper — but it builds design-and-policy obligations that sit adjacent to content moderation and carry their own liability exposure. A platform can still moderate exactly as it likes under Section 230 while simultaneously facing suit for failing to have had a "reasonable policy" addressing a listed harm category. The rule of construction protects the statute's language. It does nothing to prevent the same speech-suppressing incentives Section 230 was designed to avoid from re-entering through a parallel track.
Age Verification Is the Actual Speech Cost
The bill's practical bite falls hardest on age assurance. To know whether the "knows or should have known a minor" standard applies, platforms need a reliable way to determine who is a minor — which means checking, in practice, the age and often the identity of every user, not just younger ones. EFF's June 24 post is blunt about the tradeoff: age-verification systems "fail more frequently for people of color, people with disabilities, and trans and nonbinary people," and they convert anonymous browsing into identified browsing for the entire user base, not a protected subset of it.
The Proportionate Path
We support several elements here: banning targeted ad profiling of children, requiring AI chatbots to disclose they aren't human, and modernizing COPPA's decade-old consent thresholds. Congress should not need a duty-of-care fight to get those done, and this bill largely avoids one. But lawmakers should be honest that the "should have known a minor" liability standard functions as a soft mandate for identity verification across the entire internet, not just services aimed at kids — and that cost falls on every adult user's privacy and anonymity, whatever the Section 230 rule of construction says on paper. The Senate should narrow that standard before this becomes law, not treat the House's 267-117 margin as license to skip the harder questions.