KOSA’s Real Censorship Risk: How 50 State AGs Could Police the Internet for Everyone

ByMarcus Vale

The Censorship Risk Is Real — But It Works Differently Than Most Headlines Suggest

The Kids Online Safety Act doesn’t contain a list of banned websites. There’s no registry of forbidden topics. The censorship risk in KOSA is more indirect — and more durable — than that. The bill creates liability incentives so broad that platforms will restrict legal speech on their own, preemptively, simply to avoid getting sued.

That’s a harder problem to fix than a bad sentence in the text. It’s structural.

What KOSA Actually Requires

First introduced in 2022 by Senators Marsha Blackburn (R-TN) and Richard Blumenthal (D-CT), KOSA imposes a “duty of care” on online platforms serving users under 16. Platforms would be required to prevent and mitigate a specified list of harms: anxiety, depression, eating disorders, substance use disorders, physical violence, online bullying, sexual exploitation, and suicidal behavior.

Each item on that list is a real problem. None of them is something a platform can actually prevent.

Platforms can’t stop a teenager from feeling anxious after reading their friends’ posts. They can’t guarantee a user won’t encounter something upsetting. So in practice, rather than targeting psychological outcomes, regulators end up targeting content — and KOSA leaves the definition of actionable “harmful content” almost entirely to the Federal Trade Commission and fifty individual state attorneys general.

The Attorney General Problem

This is where the bill becomes genuinely dangerous. Sen. Blackburn, one of KOSA’s authors, has publicly described education about race discrimination as “dangerous for kids.” Dozens of states have moved to restrict classroom instruction on race, gender, and sexual identity. If KOSA passes with AG enforcement intact, platforms face a stark choice: host content that a state AG might sue them over, or pull it to avoid litigation.

Most will pull it. Lawsuits are expensive. The incentive runs in exactly one direction.

The Electronic Frontier Foundation has noted that the phrase “prevent and mitigate” is vague enough to support almost any enforcement action a motivated AG wants to bring. An AG who believes safe-sex information harms minors has a defensible argument under KOSA’s language. So does one who considers transgender health resources dangerous. Platforms can’t predict where 50 different AGs will draw the line, so they draw it conservatively — meaning well above where the law actually requires.

That gap between the legal threshold and the platform’s actual behavior is where censorship lives.

Age Verification and the Privacy Cost

To comply with KOSA, platforms need to know who is under 16. That requires age verification. And age verification at internet scale means either government ID checks or dependency on third-party identity brokers like Clear or ID.me.

Neither option is benign. Government ID checks erase anonymous speech online — a protection that matters for journalists, activists, abuse survivors, and LGBTQ+ youth in hostile households who rely on anonymity to safely seek information. Third-party identity companies collect and store biometric data or government document scans at enormous scale, creating centralized databases that are historically attractive to hackers.

The EFF’s analysis of the House-passed KIDS Act — which shares KOSA’s age-verification logic — describes this as building exactly those kinds of centralized sensitive-data honeypots. The ACLU has made the same argument: age-gating the internet doesn’t just affect minors, it forces everyone to prove who they are before they can read.

Who Actually Gets Hurt Most

The bill’s supporters frame any opposition as either pro-Big Tech or anti-child. That framing erases the groups who lose the most under KOSA.

LGBTQ+ youth in conservative states often depend on the internet to access community support, health information, and crisis resources that simply aren’t available locally. A coalition of more than 90 civil society and LGBTQ+ rights organizations formally opposed KOSA on exactly these grounds. If platforms over-censor gender identity and sexual health content to dodge AG liability in hostile states, those young people lose one of their primary support systems.

Harm reduction resources face the same threat. Needle exchange locations, naloxone guides, safe injection site information — all legal, all life-saving. Under a broadly interpreted KOSA duty of care, a state AG who views these resources as encouraging drug use has a plausible basis to sue any platform that displays them.

The ACLU put it plainly: “Censorship does not keep kids safe.”

Where Things Stand in Mid-2026

The Senate passed KOSA 91-3 in July 2024. The House never voted on it, and the bill died when the 118th Congress ended. Senators Blackburn and Blumenthal reintroduced it in May 2025 as S.1748 in the 119th Congress, where it sits in the Senate Commerce Committee.

In June 2026, the House passed a related but different bill — the Kids Internet and Digital Safety (KIDS) Act, H.R. 7757 — 267 to 117. The KIDS Act draws from both KOSA and COPPA 2.0, but it omits KOSA’s duty-of-care provision entirely. Senate co-author Blumenthal declared the House version “dead” almost immediately; Sen. Ted Cruz signaled openness to negotiations.

That leaves both chambers at an impasse. Senate sponsors want the duty-of-care language they’ve fought for since 2022. House Republicans stripped it over First Amendment concerns. Digital rights groups, including the EFF and ACLU, oppose both versions on age-verification grounds.

Whatever eventually emerges from that standoff will determine whether this becomes a real child safety law or a censorship framework with sympathetic branding. The two aren’t mutually exclusive — but right now, the bill’s enforcement architecture points more toward the latter.

Sources


Similar Posts