A Deal Eight Months in the Making
On July 20, 2026, negotiators from France's National Assembly and Senate, meeting as a commission mixte paritaire — the reconciliation committee that resolves differences between the two chambers — agreed on final text banning social media access for anyone under 15 and prohibiting smartphones in secondary schools. The bill was deposited on November 18, 2025, passed the Assembly on January 26, 2026, was amended by the Senate on March 31, and the joint committee finalized compromise language after convening on July 9 (Assemblée nationale legislative dossier). President Emmanuel Macron wants the law in force before the September school year, which gives regulators roughly six weeks to stand up an enforcement system that, on the evidence from the one country that has already tried this, does not yet exist anywhere (Reuters via Cyprus Mail).
What the Text Actually Requires
The reconciled bill does two things. It bars under-15s from platforms the government designates as dangerous by decree — Instagram, TikTok, Facebook and Snapchat are the named examples, chosen because their recommendation systems are judged to threaten children's development — while allowing conditional, parent-supervised access to other services. And it extends France's existing collège phone ban into lycées from September 2026, formalizing the "pause portable" pouch system already piloted since 2025 (MediaNama). Enforcement of the social media provision hinges on age-verification systems that platforms must deploy and that France's privacy regulator, the CNIL, must approve.
That approval requirement is the hard part, and the CNIL itself has been unusually candid about why. In its published guidance on online age verification, the regulator rejects both direct age-declaration (trivially bypassed) and blanket identity-document uploads (a privacy and data-breach risk), instead pushing a "trusted third party" model: an intermediary confirms a user clears the age threshold without revealing which site they're visiting, while the site itself never learns the user's identity. The CNIL has even built a zero-knowledge-proof demonstrator toward that end. But its own assessment is blunt: "l'évolution vers un monde numérique fermé... présente des risques importants pour les droits et libertés" — a shift toward a closed, authenticate-everywhere internet carries serious risks to rights and freedoms — and no deployed solution yet meets the bar the regulator has set (CNIL). France is legislating a September deadline for a technology its own privacy watchdog says isn't ready.
The Case for the Ban, Stated Fairly
The strongest argument for this bill isn't speculative. Algorithmic feeds optimized for engagement have measurable associations with adolescent sleep loss, body-image harm and compulsive use, and parents and teachers report — accurately — that phones in classrooms degrade attention and peer interaction. A legislature responding to that evidence with a bright-line age rule and a device-free classroom is not acting irrationally; French unions like Snes-FSU support the phone measure even while flagging patchy enforcement in collèges. Reasonable people can back the underlying goal without endorsing the implementation.
Australia Already Ran This Experiment
The implementation is where France should be paying closer attention to Australia, which enacted the world's first under-16 social media ban, effective December 10, 2025, covering Facebook, Instagram, Snapchat, TikTok, YouTube and several other platforms. Seven months in, the results are instructive rather than reassuring. Australia's eSafety Commissioner reported more than 4.7 million accounts believed to belong to under-16s removed, deactivated or restricted — a large enforcement number that nonetheless coexists with a parental survey finding nearly 70% of children who had accounts before the ban still held onto a Facebook, Instagram, Snapchat or TikTok login, and close to half still used YouTube. Facial age-estimation tools proved unreliable near the 16-year threshold, and users who declared themselves underage could simply "correct" their stated age to regain access (Tech Policy Press). Reported cyberbullying and image-based abuse complaints from underage users showed no clear decline. Canberra's response has been to double maximum penalties for platforms rather than fix the verification technology itself.
Why the September Deadline Is the Real Risk
France's bill compounds Australia's problem by adding a regulator pre-approval requirement the CNIL has signaled it isn't ready to grant with confidence, on a timeline set by a political calendar rather than a technical one. Rushing verification infrastructure that the privacy regulator itself calls unsolved invites exactly the failure mode Australia lived through: a headline compliance number, a much larger circumvention rate, and a de facto push toward the identity-document verification the CNIL explicitly warned against — because platforms racing a September deadline will reach for the crudest tool that clears the bar, not the privacy-preserving one the regulator prefers.
The European Commission is watching closely: Commission President Ursula von der Leyen has said Brussels will propose EU-wide legislation restricting children's social media access after the summer, following an expert panel report. If France's rollout produces an Australia-style compliance theater rather than durable protection, that becomes the template 27 member states inherit. A proportionate alternative exists in the CNIL's own recommendations — mandate the privacy-preserving verification architecture, give it a realistic implementation runway rather than a school-year press release, and hold platforms to a phased compliance standard tied to actual technical readiness rather than a political date. Child safety and an open internet are not opposed goals; a rushed mandate that forces either fake compliance or mass identity collection achieves neither.