KOHEN AVOCATS FIRM · PARIS 16
France’s Under-15 Social-Media Ban Falls: What the Constitutional Council’s Ruling Means for EU Platforms

Hassan KOHEN acts in urgent matters, from the police station to the assize court. Initial strategic analysis: €80 incl. VAT, personal reply within 24 hours.

100% confidential · Legal professional privilege · No obligation

Article generated by artificial intelligence, through a process designed and supervised by the firm

Source: French Court of Cassation – Judilibre and Légifrance open data

Paris Bar Police custody, investigation, assize court CNB listing avocat.fr
Maître Hassan KOHEN, avocat au Barreau de Paris
Hassan KOHEN
Attorney at the Paris Bar

France’s Under-15 Social-Media Ban Falls: What the Constitutional Council’s Ruling Means for EU Platforms

Speak to a lawyer 06 89 11 34 45

France’s Constitutional Council has delivered a sharply limited but internationally important answer to the first European attempt to impose a blanket social-media age threshold. In decision no. 2026-911 DC of 14 August 2026, it censured Article 1 of the law intended to bar children under 15 from online social-networking services. The ruling does not reject child protection, does not invalidate the Digital Services Act, and does not create a general right to access every online service without safeguards. It says something more exact: Parliament could not prohibit access across an entire category of services without distinguishing their functions, risks, protections, the child’s circumstances, or the role of parents. The Council also identified the hidden consequence of the law: because platforms would need to know who is under 15, every user, including adults, could be required to prove age before connecting.

That combination gives the decision global reach. Platforms operating from the United States, Asia or elsewhere still face European duties toward minors. The European Commission is simultaneously building a privacy-preserving age-verification framework and applying the Digital Services Act to large platforms. Yet a French measure that turns age assurance into a universal gateway must now be rebuilt around necessity, proportionality and legally defined privacy safeguards. For operators, parents and users, the practical question is no longer whether France will regulate children online. It is how far a French rule may reach into an international service, what data it may demand, and which regulator or court can be challenged when the system fails.

I. Why did France’s under-15 social-media ban fail constitutional review?

A. Why the Council treated a blanket ban as a disproportionate restriction

The starting point is not a technology policy preference but the constitutional status of online communication. Article 11 of the 1789 Declaration protects the free communication of thoughts and opinions. Its official wording calls that freedom “l’un des droits les plus précieux de l’homme” and allows restrictions only for abuses defined by law. The relevant text is available in full through Article 11 of the Declaration of the Rights of Man and of the Citizen. The Council has adapted that principle to modern communication: access to online public communication services and the ability to express oneself through them are now part of the protected space.

The Council did accept the legitimacy of the legislature’s objective. Protecting children from harmful content, manipulative design, contact with dangerous adults, harassment and other online risks can justify limiting access for minors. The decision also recognised the constitutional requirement to protect the best interests of the child and the constitutional objective of preventing threats to public order. The ruling is therefore not a declaration that platforms are constitutionally neutral or that parents and public authorities must tolerate every risk. It is a decision about the legal design of the restriction.

The defect was the law’s breadth. The contested rule could apply to online social-networking services without a meaningful distinction based on their content, functionality, design, exposure to risk or available protective measures. A service built around private family communication, a professional network, a video-sharing platform and an anonymous public forum could fall within the same prohibition even though the risks were not equivalent. The Council’s reasoning is that a legislature may pursue a strong objective, but the means must be connected to the actual risk addressed.

The decision uses a familiar constitutional test: restrictions on expression must be “nécessaires, adaptées et proportionnées à l’objectif poursuivi” — necessary, appropriate and proportionate to the objective pursued. That formula matters for every future version of the law. A new measure cannot simply repeat the under-15 threshold and move the technical burden to platforms. It will need a defensible explanation of which services are covered, what risk justifies the threshold, why less restrictive measures are insufficient, and how the rule responds to differences in age, maturity and context.

The Council also focused on the absence of any individual assessment. The law treated every minor below 15 as presenting the same regulatory problem. It did not allow the risk analysis to account for the child’s age, degree of maturity, family situation or the nature of the service. Nor did it define conditions under which a parent or legal representative, informed about the risks and protections of a service, could lift the prohibition, narrow it or authorise access to a specific service in the child’s interests.

That reasoning does not mean that parental consent will necessarily be enough in a replacement statute. Parental authority is itself structured by law. Article 371-1 of the Civil Code describes parental authority as “an ensemble of rights and duties” directed toward the child’s interests, safety, health, privacy, morality, education and development, while requiring that the child be involved according to age and maturity. The constitutional lesson is that Parliament must decide how that authority interacts with a platform ban; it cannot silently remove the question from the legal framework.

The immediate consequence is narrow but decisive. Article 1 cannot be promulgated in the form examined by the Council. The rest of the statute is not automatically erased merely because its central access prohibition fell. Any remaining duties must be read separately, and any new administrative implementation must have a valid legal basis. The decision is not a licence for platforms to abandon every child-safety measure. It is a warning against treating a universal age gate as a substitute for risk regulation.

The international audience should also understand the institutional point. This was a prior review of legislation, not a damages judgment against a platform and not a ruling on the legality of a particular algorithm. Under Article 61 of the Constitution, sixty deputies may refer a statute to the Council before promulgation, and the Council must rule within the constitutional procedure. The decision itself therefore answered whether Parliament had adopted a constitutionally sufficient rule; it did not decide whether TikTok, Meta, Discord or a smaller service complied with a separate European or data-protection obligation.

B. Why universal age verification became the law’s second constitutional weakness

The most consequential part of the ruling concerns adults. A prohibition directed at minors under 15 sounds targeted, but a platform cannot reliably prevent a minor from opening an account unless it has a method to establish age. The Council observed that the law therefore implied that every person, including an adult, would have to prove age before accessing the affected services. The practical burden would not have been limited to children who triggered a risk signal. It could have become a general identity or age checkpoint for the entire French user base.

That design raised a separate privacy problem. The Council relied on the constitutional protection of private life, which it connected to Article 2 of the 1789 Declaration. The text of Article 2 of the Declaration protects natural and imprescriptible rights including liberty, property, security and resistance to oppression. The Council’s settled constitutional reasoning derives the right to respect for private life from that protection. Article 9 of the Civil Code states the ordinary-law rule in simpler terms: “Chacun a droit au respect de sa vie privée.” The full provision, including the judge’s power to order measures to stop an infringement, appears on Article 9 of the Civil Code.

The problem was not that age assurance is always unlawful. The problem was that Parliament did not determine the conditions and limits of the proof demanded from users. The statute did not sufficiently identify what data could be collected, who would process it, how long it could be retained, whether the platform or an independent intermediary would receive the information, how errors would be corrected, or how adults could access the service without disclosing more identity information than necessary. Those omissions were not minor technical details. They were the safeguards that would make the restriction compatible with private life and free communication.

The distinction between age proof and identity proof is central. A privacy-preserving system could confirm that a user is above or below a threshold without disclosing a name, address, identity document or browsing history to the platform. A system that requires an image of a passport, a facial scan or a bank-account match may create a much larger pool of sensitive information. The Council’s decision did not approve any one technical model, but it makes clear that Parliament must legislate the safeguards rather than leaving the most intrusive choices to private operators.

French data law reinforces that allocation of responsibility. Article 8 of the amended Data Protection Act of 6 January 1978 identifies the CNIL as the national supervisory authority under the General Data Protection Regulation and assigns it information, supervision and enforcement functions. Article 20 of the same Act allows the CNIL president to warn a controller or processor when planned processing may breach the GDPR or the French Act, and to issue a formal notice where compliance can be achieved. The statutory design therefore expects a legal framework in which the data regulator can test the system; it does not make privacy a box to be checked after deployment.

There is also a criminal-law backstop, although it is not a substitute for a lawful design. Article 226-16 of the Criminal Code punishes certain unlawful personal-data processing with five years’ imprisonment and a fine of €300,000. A platform should not read that provision as a prediction that every defective age-assurance trial becomes a criminal case. It is a reminder that a controller cannot treat a large-scale identity experiment as an ordinary product feature without mapping its legal basis, security, retention, access and accountability obligations.

The Council’s privacy reasoning also closes an easy political response. Parliament cannot say that platforms will “work out the details” after the ban has been enacted. Details that determine whether an adult can participate, whether a child can appeal a false classification, and whether an intermediary can link age credentials to identity are part of the restriction itself. They require legislative guarantees. A future law will need a data architecture, not just an age number.

The result is a constitutional boundary with cross-border consequences. A French platform may be supervised by the CNIL, but a foreign platform may process an age credential through entities, cloud infrastructure and verification providers in several countries. The more the French rule depends on international processing, the more precisely it must define controller roles, transfer safeguards, audit rights and remedies. A French user cannot be told that a constitutional privacy guarantee disappears because the service’s database sits outside France.

II. What does the ruling mean for EU and international platforms?

A. Does the French ruling cancel European duties toward minors?

It does not. The Digital Services Act remains directly applicable across the European Union. Article 28 requires providers of online platforms accessible to minors to put in place appropriate and proportionate measures to ensure a high level of privacy, safety and security for minors. The complete instrument is available in the official Digital Services Act text on EUR-Lex. The French decision is consistent with the words “appropriate and proportionate”: the Union rule also rejects a one-size-fits-all assumption, even though it imposes a real duty on services to assess and mitigate risks.

The DSA and the French statute operate on different legal planes. The DSA imposes platform duties, supervision and systemic-risk controls. The French statute attempted to create a specific access prohibition for people under 15. The Constitutional Council held that the French prohibition was constitutionally defective; it did not hold that the DSA’s child-protection obligations were defective. A platform cannot use the French ruling as a reason to remove parental controls, safety-by-design measures, reporting channels, or risk assessments required by Union law.

The European Commission’s current age-verification work makes the distinction especially important. Commission Recommendation (EU) 2026/1035 describes a common framework for EU-wide age-verification technologies and encourages solutions that allow age to be established without unnecessary disclosure of identity. The official recommendation is available through EUR-Lex. The Commission’s policy page also explains its blueprint and privacy-preserving approach at The EU approach to age verification.

That framework does not automatically rescue a French blanket ban. A Commission recommendation is not a French constitutional amendment and does not supply every statutory guarantee required by the Council. Conversely, the fact that France’s Article 1 was censured does not prevent EU law from requiring platforms to use age assurance in particular risk settings, where the measure is appropriate and proportionate and has a valid legal basis. Operators must analyse the two regimes together rather than choosing the one that appears less demanding.

The French law’s use of EU definitions was another reason the case mattered beyond France. Current Article 6 of the Law for Confidence in the Digital Economy defines a “platform online” by reference to Article 3 of the DSA and a “service de réseaux sociaux en ligne” by reference to the Digital Markets Act. The definitions can be checked in Article 6 of Law no. 2004-575. That drafting technique creates legal interdependence: a French legislature relies on Union concepts, but the constitutional court still reviews the national consequence imposed on users and platforms.

For very large online platforms and search engines, supervision may also be divided. The European Commission handles core DSA responsibilities for designated very large services, while national Digital Services Coordinators and other authorities retain roles within the Regulation. The CNIL remains relevant to personal-data processing, and French administrative or judicial courts may hear disputes arising from national implementation. A foreign provider should map the competent forum before sending a challenge to the wrong regulator.

The international question is territorial reach. A French rule may regulate access offered to users in France, even where the provider is incorporated elsewhere, but it does not become a universal rule for every user on the same global service. The operator must decide whether to build a France-specific flow, an EU flow or a global flow. A global flow may reduce engineering complexity, but it can impose unnecessary data collection on users outside France and create conflicts with other national laws. A France-specific flow may be legally narrower but harder to operate and explain.

The Court of Justice’s case law supplies a useful warning. In Case C-507/17, Google LLC v CNIL, the Court rejected an automatic obligation to remove search results worldwide, while requiring effective protection within the Union. The official case is available on CURIA. The comparison is not exact: the French case concerns access and age assurance, not delisting. The principle is nevertheless valuable. A national or European remedy can have strong local effect without automatically becoming a worldwide command, and a platform should assess territorial scope rather than applying every European restriction globally by default.

Data transfers add a second international layer. Age credentials, device signals and fraud-prevention records may move between the European Economic Area and third countries. The Court of Justice’s decision in Case C-311/18, Data Protection Commissioner v Facebook Ireland and Schrems, is the official reference point for transfer safeguards and the need to assess protection in practice; the judgment is published on CURIA. A provider cannot solve a French constitutional problem by exporting the age database to a jurisdiction where the user has fewer effective rights.

The decision therefore leaves a workable route for platforms, but not a shortcut. A service can combine risk-based design, privacy-preserving age assurance, child-safety settings, independent audits and meaningful appeals. It must be able to explain why the method is necessary, what it proves, what it does not prove, which entity controls the data, and how a user can obtain redress. The next French statute will be judged against that level of concrete architecture.

B. What should platforms, parents and users do after the decision?

Platforms should begin with a legal and technical inventory rather than simply turning off every age check. They should identify the services and features accessible to minors, distinguish public discovery from private communication, document risks by age group, and record which protections are already available. That evidence can support a proportionate design under Article 28 of the DSA and can show French authorities that the provider is addressing the risk the legislature intended to target.

The next step is data minimisation. A provider assessing an age-assurance vendor should ask whether the platform receives only a yes-or-no age attribute, whether the vendor deletes the underlying evidence, whether the credential can be reused without a new identity disclosure, whether biometric processing is involved, and whether the system can be audited without exposing children’s data. The answer should be contractually enforceable. A vague claim that a supplier is “privacy friendly” is not a substitute for a controller-processor analysis, a data-protection impact assessment where required, security controls and retention limits.

Users should not assume that every request for an identity document is now mandatory in France. The censured Article 1 cannot create the blanket access prohibition in the form reviewed by the Council. If a platform requests age information for another lawful reason, the user should be able to understand the purpose, legal basis, categories of data and retention period. The request should distinguish proof of an age threshold from proof of identity. A refusal may have consequences for that specific service, but the provider should be able to explain the legal and contractual basis rather than invoking the old statute as if it had survived unchanged.

Parents retain an important role, but parental authority is not an unbounded override. The legal framework expects parents to protect the child’s safety, health, privacy and development, while taking account of age and maturity. A redesigned law could permit parental involvement for certain services, require stronger safeguards for others, or prohibit particular high-risk features regardless of consent. The constitutional defect was not that Parliament protected minors; it was that Parliament removed all meaningful differentiation from the decision.

A parent who receives a false age classification should preserve the notice, screenshots, account information, verification steps and correspondence. The first remedy is usually the platform’s internal review or complaint process. A data-protection complaint may then be directed to the competent supervisory authority, especially where the dispute concerns excessive collection, inaccurate data, opaque automated classification or an inability to exercise access and correction rights. A French lawyer can help determine whether the dispute belongs before the CNIL, a civil court, an administrative court or a DSA authority.

For platforms facing a future implementing decision, the litigation path will depend on the act. A general statute may be challenged through the constitutional mechanisms available at the relevant stage; an administrative decision, guidance document or enforcement measure may call for a claim before the competent administrative court, potentially accompanied by an application for interim relief. A contractual suspension or account closure may instead require civil proceedings. The Constitutional Council’s decision is not an appeal route against every future platform measure. Its value is as a controlling legal standard: the French administration and Parliament must respect the distinction between a legitimate objective and an insufficiently designed restriction.

The binding effect is clear. Article 62 of the Constitution provides that a provision declared unconstitutional under Article 61 cannot be promulgated or implemented and states that constitutional decisions are not subject to appeal and bind public authorities and courts. The official French formulation is short: “Les décisions du Conseil constitutionnel ne sont susceptibles d’aucun recours.” A platform cannot seek a second appeal against the decision, but Parliament can legislate again within the constitutional boundaries identified by it.

The likely legislative response is not difficult to predict, although its precise form remains open. Parliament could narrow the covered services, create a risk classification, allow exceptions or parental intervention under defined conditions, impose a separate age-assurance intermediary, and assign responsibilities between platforms, providers and public authorities. It could also provide an appeal mechanism for age-assurance errors and a prohibition on retaining identity documents once the threshold is verified. Each choice would need to be tested against the DSA, the GDPR, French data law and the constitutional requirements of free communication and private life.

International operators should prepare for legal divergence. Australia, the United Kingdom, individual EU Member States and other jurisdictions may adopt different age thresholds, definitions and verification methods. A service that treats France’s 15-year threshold as a global setting may expose users elsewhere to an unnecessary restriction. A service that treats the French ruling as permission to ignore all age controls may breach European duties. The safer approach is a jurisdiction matrix that records the threshold, covered service, controller, verification method, retention rule, appeal channel, regulator and territorial scope for each market.

The ruling also creates a media and policy lesson. The visible controversy was whether children should use social media. The legally decisive question was how the state could make a private platform police age at scale without defining the data safeguards and without assessing the different risks of different services. That is why the judgment matters to international counsel, compliance officers and investors. The outcome may shape the design of age assurance far beyond France, particularly as the European Commission’s common framework moves from pilot technology toward real deployment.

Conclusion

Decision no. 2026-911 DC does not end France’s effort to protect minors online. It ends one legislative technique: a general under-15 ban applied across social-networking services, backed by an age-verification obligation whose privacy conditions were left undefined. The Constitutional Council accepted the objective but required the means to be necessary, appropriate and proportionate. That requirement has practical content. Parliament must distinguish services and risks, decide how parents and children participate in the decision, and define the safeguards for every user whose age is checked.

For European and international platforms, the message is two-sided. The French rule cannot be used as an unbounded global mandate, but the judgment does not suspend the DSA, CNIL supervision or the European movement toward privacy-preserving age assurance. Providers should document risk, minimise data, separate age proof from identity, secure cross-border processing and preserve a real route of appeal. Users should ask what is being verified, by whom and for how long. A revised French law can survive if it answers those questions in the statute itself.

📄 Official circular