KOSA Clears the Senate Committee and Heads for a Full Vote
The Kids Online Safety Act cleared the Senate Commerce Committee unanimously on 5 August 2026 and now heads to a full Senate vote. KOSA never uses the phrase age verification, and that is precisely why it matters: the obligation it creates only works if a platform knows which of its users are minors, and knowing that at scale means checking everyone.
The same markup produced an irony worth recording. The bill that does spell out age checks in its text, the SCREEN Act, won its vote and still failed to advance.
What the Senate Commerce Committee did on 5 August
| Bill | Outcome |
|---|---|
| Kids Online Safety Act (KOSA) | Advanced unanimously, goes to a full Senate vote |
| SCREEN Act | Favourable vote 15-13, but not advanced: the committee lacked a quorum |
| Youth AI Privacy Act | Advanced out of committee |
| CHATBOT Act | Advanced out of committee |
The quorum detail is the part most coverage skipped past. Enough senators voted in favour of the SCREEN Act, but too many of those votes were cast by proxy and too few members were physically in the room, so the tally could not carry the bill out of committee. Procedure, not opposition, stopped it.
Why KOSA implies age checks without naming them
KOSA's core mechanism is a duty of care: a covered service must take reasonable steps to avoid exposing minors to categories of content the law treats as harmful. Read that as an engineer rather than a lawyer and one requirement falls out immediately. To apply a different standard to minors, the service has to know who the minors are.
There is no way to know that about a stranger without asking. In practice, as the Electronic Frontier Foundation put it on 3 August, the pressure lands on age verification or age estimation systems, and those systems mean "checking government IDs, performing facial analysis, checking your bank records, or collecting other personal information". None of that is limited to children, because you cannot sort minors out of a crowd without inspecting the whole crowd.
The second effect is on speech rather than data. Faced with liability for content a regulator later calls harmful, the cheapest defence is to remove the category. EFF's summary of that dynamic is blunt: companies "shut down potentially controversial forums in order to avoid legal action from government bureaucrats".
Four bills, one direction
KOSA did not move alone. The Youth AI Privacy Act and the CHATBOT Act advanced the same day, and the SCREEN Act would have if the room had been fuller. EFF, which opposed all four, described the common thread the day before the markup as an attempt to "age-gate the internet and block young people from speaking and accessing lawful speech online".
We covered the SCREEN Act in detail when its text turned out to name proxy network IP addresses directly: see The SCREEN Act Puts VPN Traffic in the Text of a Federal Age-Check Bill. That bill stalled on attendance this week, but its approach did not lose an argument, and nothing prevents the committee from taking it up again with a full room.
What age checks mean for VPN users
Nothing changes today: a committee vote is a step, not a law, and KOSA still needs the full Senate. What the vote does is confirm the direction of travel, and the pattern from states that already run age checks is consistent. When Missouri's law took effect, interest in tools that change the apparent location of a connection rose sharply, because a check tied to where you appear to be is a check you can sidestep by appearing somewhere else.
It is worth being precise about what that does and does not solve. Moving your apparent location can put you outside a jurisdiction's checks; it does nothing about an account that already holds your verified identity, and nothing about a platform that decided to remove a category of content for everyone. The privacy question raised by these bills is not only who can reach a site, it is how much documentation the internet asks for before it lets anyone speak.
