A notebook entry to begin: the internet was built by people who did not ask your age, and for thirty years that was treated as a feature. Mexico’s decision to require digital platforms to shield minors from harmful content — covered by Reuters — is one of several moves worldwide converting that feature, in regulators’ eyes, into the bug. What follows is an attempt to record the age-assurance debate as it actually stands: the technologies on offer, the objections that matter, and the questions no capital has yet answered well.
Our notebook notes have followed parallel laws in Australia and Europe; Mexico’s version arrives with the technology in a noticeably more mature state, which changes the argument’s terms.
Entry one: the methods, ranked by intrusiveness
Self-declared birthdates — intrusive not at all, effective barely — are being regulated out of credibility. Document uploads work but concentrate identity papers on platforms never designed as vaults. The frontier is twofold: estimation, where a camera image is converted into an age band and discarded, and attestation, where a wallet or provider confirms a single fact — over 16, over 18 — and nothing else. The policy fight is increasingly between these two frontiers, not between checking and not checking.
Accuracy is the frontier methods’ weak flank. Estimation errs by a few years in either direction, errs unevenly across faces and lighting, and errs worst exactly at the boundary ages the law cares about. Every system therefore needs an appeal route for the wrongly excluded — which quietly reintroduces the documents the system was built to avoid.
Entry two: the objections that deserve ink
The serious privacy objection is not ‘checks are annoying’. It is that age gates for children become identity gates for adults: to prove some users are young, systems must test all users, normalising a checkpoint between citizens and lawful speech. Add breach risk — a stolen age-credential database is a gift to fraudsters — and exclusion risk for the undocumented, and the objection is substantial. The serious protection objection replies: the ungated status quo delivers children to algorithmic systems optimised without them in mind, and ‘do nothing’ is also a decision with victims.
Noted, also, the displacement problem: determined teenagers migrate to unregulated corners, a dynamic every prohibition regime in history would recognise.
Entry three: what good would look like
A defensible settlement is visible in outline across the better drafts in several countries: privacy-preserving attestation as the default; data minimisation written as law, not promised as policy; independent accuracy and bias audits; no platform-held identity documents; and parental tools treated as complements rather than substitutes. Whether Mexico’s implementing rules land near that outline is the detail worth watching — implementation, as ever, is where values become settings.
Closing note
What this notebook will watch
Four details will tell us how the age-assurance era actually settles. Whether estimation systems publish accuracy figures broken down by age and ethnicity, or hide behind averages. Whether any major platform is fined for retaining check data it promised to delete — the moment the privacy clauses acquire teeth. Whether teenagers’ migration to unregulated spaces shows up in measurement, and how honestly officials respond when it does. And whether the attestation model — prove the fact, not the identity — survives contact with law-enforcement requests for more. Mexico’s implementation will supply early answers to all four. They will be recorded here as they arrive, with the dates attached, because in this debate the sequence of decisions matters as much as the decisions themselves.
One more consideration belongs in the record: parents are not a single constituency in this debate. Some want platforms to act as strict gatekeepers; others want tools, information and the final say left at home. Laws that pretend otherwise — that write one family’s preference into everyone’s browser — tend to satisfy no one for long, which is why the design details, not the speeches, will decide how Mexico’s experiment is judged.
Age assurance is unglamorous plumbing that will quietly restructure who can reach what online. The debate deserves better than its loudest versions on both sides, and this journal will keep its notes in the open — start from our homepage for the running file.
