On 14 September 2026, Emmanuel Macron said on X that, "after rigorous technical work," the government was notifying the European Commission of a reworked bill banning social media for under-15s (Khaleej Times/AFP). The notification is the EU-law compliance step. It does not answer the objection that sank the first law, which was constitutional.
The case for the ban deserves a fair hearing
The strongest argument for an age floor is that children face real, documented pressures online, and that platforms have had years to fix this voluntarily. Individual parents cannot out-negotiate an engagement-optimised product. A statutory minimum age gives them a clear rule and gives teenagers cover to say no, as peers do. The Constitutional Council accepted part of this reasoning: it recognised that protecting children is a legitimate aim that can justify limiting minors' access to online services.
The Council still rejected the law. In Decision n° 2026-911 DC of 14 August 2026, it found Article 1, the ban itself, contrary to the Constitution (Conseil constitutionnel).
What the Council actually said
The decision is narrower and more useful than the headline suggests. It rests on two findings.
- Freedom of expression. The Council held that the right to free communication includes the freedom to access online services and to express oneself on them. The ban was not tied to the features, content or risks of each service. Its exceptions were narrow. It gave parents no way to consent to or limit access, and it allowed no assessment of the individual minor. The Council concluded that the restriction was not "adaptée, nécessaire et proportionnée" (para. 18).
- Privacy. A ban for under-15s means everyone, including adults, must prove their age. Parliament set neither the conditions nor the limits of that proof. The Council found it had not provided the legal guarantees needed to protect privacy (paras. 19–21).
Press reports add that the ban could reach services whose risks to minors' health and safety had not been established (The Next Web). The adopted text defined covered services by reference to the 2004 digital economy law's notion of a social network and named no platform. It left age verification to platforms, subject to approval by the privacy regulator CNIL, and specified no method. Platforms were to stop creating new under-15 accounts from 1 September 2026, with a transition for existing accounts to 1 January 2027.
Notification is necessary, not sufficient
The government's reply to this ruling was to notify Brussels. That is a sensible step, since national rules on online services must sit within the EU framework. But it addresses a different question from the one the Council asked. The Commission can check whether a French rule fits EU law. Only French drafters can fix a law the Council found disproportionate.
The EU context matters here. The Commission's July 2025 guidelines on protecting minors under the Digital Services Act recommend full age verification where national rules set a minimum age for certain services such as defined categories of social media. They recommend less intrusive age estimation in other cases (European Commission). The guidelines are voluntary. They also treat age assurance as something that must be "accurate, reliable, robust, non-intrusive and non-discriminatory", which is close to the privacy proportionality test the Council applied. A French law that satisfies one test is better placed to satisfy the other.
What a proportionate rewrite looks like
We support a rewrite and think the Council's reasoning shows how to do it. Four design choices follow from the ruling.
- Define covered services by demonstrated risk. Feeds built on infinite scroll, algorithmic amplification and stranger-contact features are a different category from messaging, encyclopedias or educational tools. A blanket category invites exactly the overbreadth the Council found.
- Build in parental and individual calibration. The Council noted the absence of any route for parental consent or case-by-case assessment. A model with consent tiers, as in some US state laws, would be easier to defend, though the US experience carries its own warnings.
- Legislate the age-check limits. Parliament, not platforms or a regulator's later guidance, must set what data may be collected, how long it is kept, and what is prohibited. Privacy-preserving, device-level or wallet-based proofs are the natural direction.
- Measure outcomes. A sunset clause and a published evaluation would turn a blunt instrument into an evidence-based policy.
The risk of exporting the problem
The American picture shows what a poorly bounded law can do. On 23 October 2026, the Electronic Frontier Foundation reported that the 11th Circuit had allowed Florida and Georgia to begin enforcing laws gating social media for young users. It argued these laws burden the speech rights of minors and adults alike through age checks (EFF). The EFF post is advocacy, not a court opinion, but its core warning applies in France: age gates are identity gates for everyone.
Macron has also asked Brussels for a bloc-wide ban. If that happens, the European Parliament and Council would be writing a rule that touches every member state's constitutional order. The French ruling is a free preview of the objections. A Union-wide ban that repeated the first French text's defects would face challenge in Luxembourg instead of Paris.
Bottom line
France has a legitimate goal, a clear judicial roadmap and a political deadline: Macron wants the ban in force before the 2027 presidential election. Deadlines are when proportionality gets sacrificed. The better path is a narrower law aimed at demonstrably risky services, with statutory privacy safeguards and parental input. It would protect children and still be defensible in court. The Council did not say that France cannot act. It said France must act precisely.