On 14 October 2026 the German federal cabinet adopted outline guidelines (Eckpunkte) for a national transitional rule on children's social media use. Children under 13 would generally be barred from social networks, with exceptions for services shown to be child-appropriate. Teenagers aged 13 to 18 would face graduated restrictions. The federal government's own summary says the rule would apply only until the EU sets a minimum age (Bundesregierung). It is not law. Family and education minister Karin Prien aims to bring a reform of the Jugendschutzgesetz to cabinet this year, and the start date is open.
The strongest case for acting
The case for the plan is serious. Platforms have run an enormous experiment on children for fifteen years, and their own minimum-age rules are widely ignored. Self-declared birthdates are not a control. Parents face a collective-action problem: a child who is the only one off a platform pays a social cost. A legal floor moves that burden from families to companies. It is also modest. The age of 13 is already the threshold most large platforms claim to enforce, so the rule mostly asks companies to do what their terms of service say.
Germany did not improvise it either. An independent expert commission presented 56 recommendations in September, and the cabinet outline draws on them (Bundesregierung).
Why a blanket bar is still the wrong tool
The commission's own chair, Olaf Köller, warned that it falls short to think only about age restrictions, according to News4teachers. A twelve-organisation alliance, including Bitkom, the national student conference and the national parents' council, argues that young people should not be excluded from digital spaces across the board. It asks instead for stricter safety obligations and better enforcement of existing EU rules. Bitkom's chief executive Bernhard Rohleder said the government is overshooting the target with a flat ban. Left-party leader Heidi Reichinnek has called the plan symbolic policy.
The free-expression objection is not a footnote. Children have speech and information interests: finding peers, news, and communities that are absent from their town. A categorical ban treats a twelve-year-old reading a science channel and one in a harassment-heavy group chat identically. The exception for demonstrably child-appropriate services is the right instinct, but the outline does not say who certifies appropriateness, by what test, or on what timeline. Until it does, the exception is a promise rather than a rule.
A parallel fight in the United States shows the risk of design-based restrictions that reach speech itself. In Meta v. Bonta, EFF argued in September that California's SB 976 burdens teens' own First Amendment rights to receive recommended content. It said narrower tools, such as limits on data collection and tracking, can protect minors with less speech cost (EFF). Germany has no First Amendment, but the proportionality test in German and EU fundamental-rights law asks a similar question: is there a less restrictive route to the same goal?
Enforcement is the real weak point
Age verification is planned from 2027 through the German implementation of the EU Digital Identity Wallet, called "d-you". The design goal is to prove age without handing over name or birthdate. That is the privacy-preserving approach, and it is better than uploading ID documents to each platform.
But the plumbing is not ready. News4teachers reports that, according to the 2026 eGovernment-Monitor, the online ID function is usable for only 44 percent of those entitled to it. The eID is also available only from age 16, so it cannot be the credential for the very children the rule targets; parents' credentials or a different attestation route will be needed. The outline does not settle which. If a ban takes effect before a trusted, low-friction age signal exists, platforms will fall back on face scans and document uploads. Adults will then pay a privacy price for a child-safety rule.
The EU layer matters more than the national one
Berlin wants the EU KIDS Act to include an under-13 limit, a ban on addictive design features, and recommendation algorithms switched off by default. The Commission proposed the act on 17 September. It would bar under-13s from social media, give 13-to-15-year-olds parent-managed "mini accounts" with limited features and a one-hour daily cap, and let people open their own accounts from 15. Services used by minors would have to limit features such as infinite scrolling, reward mechanics and sleep-hour notifications by default, and very large platforms would have to demonstrate that their services are safe for children (European Commission).
This exposes a mismatch. Germany's national stopgap sets a graduated regime for 13-to-18-year-olds that may differ from the Commission's 15-year line. A national rule that runs ahead of a pending regulation also risks conflict with the EU's single-market logic. If the final EU text lands at a different age or mechanism, companies will have built two compliance systems in a year.
What a proportionate version looks like
First, publish the exemption test before the ban. Define child-appropriate by objective criteria: no algorithmic amplification of strangers' content, no open direct messaging from unknown adults, no behavioral advertising. Services that meet it should qualify automatically rather than by case-by-case approval.
Second, regulate design, not just access. Defaults that turn off autoplay and engagement-optimized recommendation for minors address the documented harm mechanism and keep speech available. The age-based gate can then be narrower.
Third, do not start enforcement until the wallet works. A rule that cannot be checked without collecting more data on everyone is a rule with a hidden cost. Sequence the law to the credential, with a sunset tied to EU adoption so that the "interim" label has teeth.
Germany is right that children deserve more than a click-through birthday box. The question for the coming cabinet decision is whether the final text builds an enforceable, narrow safeguard or a headline that critics will rightly call symbolic.