Age verification on social media: why adults get caught in the net too

Banning under-15s forces (almost) everyone to prove their age. The Constitutional Council’s privacy critique, ID upload vs estimation vs parental controls, Australia, EU graduated access — decode the tradeoff, not a bypass guide.

To stop a 14-year-old from opening TikTok, you often have to ask a 34-year-old to prove they are not 14. That is not a messaging glitch: it is the cold logic of an age floor. France’s Constitutional Council spelled it out on August 14, 2026 when it struck down the under-15 social media ban — including for lack of legal safeguards around age proof.

This piece is about the privacy trap, not the full ruling (see what the Council struck down). It is not legal advice and not a how-to evade rules.

The paradox in one line

A ban aimed at minors does not stay with minors. Once access is refused under an age threshold, the service must tell “too young” from “old enough.” Without proof, anyone can click “I am 18.” With proof, every person — including adults — enters the verification tunnel.

The Council puts it plainly (decision No. 2026-911 DC, as reported by Numerama / L’Express): barring under-15s implies, by itself, that every person, even an adult, must prove their age before access. That is not alarmism. It is the mechanism.

What the French law left unlocked

The censored version told platforms to deploy one or more age-check systems at sign-up, without fixing in statute:

  • what may be collected (selfie, ID scan, facial age estimate, parental attestation…);
  • how long data is kept;
  • who can see it (platform, contractor, state);
  • which limits and penalties apply if data leaks or is reused.

The judges conclude that, without setting conditions and limits, lawmakers failed to provide the legal safeguards privacy requires. Useful angle: you can want to protect children and refuse to turn age proof into a mass identity file — the two goals do not cancel each other.

Three families of methods (and their frictions)

Not a shopping list for “the best one”: a schema for the tradeoff.

Approach Idea Typical friction
ID upload Scan / read a national ID, passport, or licence Strong age assurance, large data surface (breach, retention, subcontractors)
Age estimation (face, behaviour…) “You look older / younger than X” Less paperwork; errors, bias, false positives / negatives; still biometric or proxy data
Parental controls / consent Child account under an adult’s responsibility Less “state vs teen,” more household dependency; unequal devices / attention

None is magic. ID upload reassures on age fraud and worries adults about privacy. Estimation sometimes reduces document scans but does not erase error or sensitive processing. Parental controls move the lever — they do not erase “how does the service know who is who?” once a hard ban applies to everyone under a threshold.

Australia, EU: same dilemma, other calendars

Australia (under-16 ban, late 2025, as France 24 notes): press reports the existence of workarounds (accounts opened in adults’ names, VPN use) and uneven usage effects by age band. That observation is not a how-to: it is a reminder that an age floor alone solves neither tech nor motivation. The stricter the check, the more pressure on legitimate users’ data.

European Union: the Commission has floated “progressive and graduated” access rather than a single wall; an expert committee notably recommended a floor around 13, with each country able to set its own (France 24 framing). The DSA (Digital Services Act — Vie publique) already piles duties on very large platforms (minors’ protection, systemic risks). “Graduated” and “DSA” do not say how to verify age without trampling privacy: they say France’s 2026 fight is not isolated.

In France, after the strike-down, the executive aims at a legally robust rewrite by spring 2027, accounting for the European frame (ruling decode). The privacy knot remains: who proves what, to whom, with which limits.

What you can / cannot say

You can say

An under-15 ban, in practice, forces age proof for (almost) all users, adults included. The Council censored the lack of legal safeguards on that point (2026-911 DC). ID upload, estimation, and parental controls have different risk / error profiles. AU and the EU show the same dilemma in other shapes.

You cannot say

“Protecting minors = inevitably filing every adult with no limits” (that is exactly the hole the judges refused). “The ruling kills all age verification” (no: it demands a frame). “A VPN fixes the issue” (neither advice nor public policy; press notes workarounds, not a solution).

Bottom line

The anti-under-15 net is not an invisible filter on teens only: it is a gate where everyone must show they may pass. The Constitutional Council struck down a law that imposed that logic without writing privacy safeguards. What comes next — France 2027, EU graduated access, Australian experiments — will turn on the proof method, not only on the number 15 or 16. Protecting children and limiting adult identity collection: one debate, two faces.

Going further

Sources

Found an error? Email us — we correct factual mistakes and note significant updates on the article. Contact us

Keep exploring