A Ban Dies on Its Way to the Schoolyard
France's Parliament adopted a law on July 21, 2026 barring anyone under 15 from social media platforms, timed to take effect at the start of the September school year. It never got there. On August 14, the Conseil constitutionnel — France's highest constitutional authority — struck down Article 1, the law's central prohibition, in Decision No. 2026-911 DC. The government still promulgated the rest of the statute as Law n°2026-813 on August 24; the age ban and the verification regime built to enforce it are simply void. President Macron has now tasked Prime Minister Sébastien Lecornu with drafting a replacement, targeted for spring 2027, according to Tech Policy Press.
The Case Parliament Was Answering
Give the ban's authors their due. French lawmakers pointed to real, parent-reported harms — anxiety, sleep disruption, harassment — associated with early social media use, and public opinion was squarely behind them: 74% of French parents supported a ban before age 15, according to a Toluna Harris Interactive poll of 1,000 parents for the Observatoire de la parentalité numérique, reported by ICI in early 2026. A platform-agnostic age floor is also administratively simpler than case-by-case content review, and France was not alone — Austria, Greece and Spain are advancing comparable proposals, per Tech Policy Press, while the European Commission has a bloc-wide position on child online safety due by December 2026. A national regulator moving first, loudly, is a defensible way to force that EU conversation.
Why the Court Said No
But wanting a real problem solved does not make any solution constitutional, and the Council's objections were structural, not cosmetic. It grounded its ruling in Article 11 of the 1789 Declaration of the Rights of Man (freedom of expression and communication) and Article 2 (privacy), and it found the ban failed on three counts, as detailed in the Council's own press release:
- No differentiation by risk. The prohibition covered every social media and video-sharing service alike, regardless of whether a platform's actual features posed the harms Parliament cited.
- No individual assessment. A mature 14-year-old and a vulnerable 10-year-old faced an identical bar, and parents had no statutory path to authorize access based on their own child's circumstances.
- Unsafeguarded age verification. The law required platforms to verify age — meaning adults, too, would have to prove they were over 15 to use ordinary social media — without specifying what data could be collected, how long it could be kept, or what would stop it from being repurposed.
The Council was explicit that it was not rejecting the objective of child protection; it ruled the measure was neither "necessary, adapted nor proportionate" to that objective. That is a narrower and more useful finding than a straightforward defeat for regulators: it is a instruction manual for what a compliant law needs to contain.
The Verification Trap
The age-verification failure is the part other jurisdictions drafting similar bans should study hardest. Any effective under-15 rule needs some way to confirm a user's age, and every current method — ID-and-selfie checks via a trusted third party, algorithmic age estimation from a photo, or an anonymized "age token" — moves personal data through new infrastructure that did not exist for this purpose. France's own data regulator, the CNIL, had previously pushed the audiovisual regulator Arcom to build "double-anonymity" into its verification framework for adult-content sites under the 2024 SREN law, precisely so no single party could link an identity to a browsing record. The under-15 ban's drafters didn't import that discipline: they mandated the check without writing the safeguards into the statute. A court did not need to speculate about hypothetical privacy harm — the law's own text left the door open.
A Law People Already Doubted Would Work
There's a deeper problem the ruling doesn't reach but the government should: even parents who wanted the ban didn't think it would hold. In the same Toluna Harris Interactive poll, 85% of French parents said the restriction would be easy for kids to circumvent — a rare case of a law's own intended beneficiaries pre-registering skepticism before it ever took effect. VPNs, borrowed adult IDs, and third-party accounts don't require a constitutional ruling to defeat an age gate; they just require a bored teenager.
The Narrower Path Still Open
None of this forecloses regulation — it specifies its shape. A statute that lets parents authorize access case-by-case, distinguishes platforms by demonstrated feature-level risk (autoplay, algorithmic feeds, direct messaging with strangers) rather than by category, and legislates the privacy safeguards for age checks rather than delegating them to an unwritten decree would likely survive review the current law did not attempt to pass. That is more legislative work than a blanket ban, and slower than a September deadline. But Lecornu now has a spring 2027 runway and an unusually specific opinion from the Council telling him exactly what to fix. The temptation, with an EU-wide child-safety framework also due by December, will be to move fast and reach for the ban again. The Council has already shown what happens then.