The future according to AI

Age verification becomes a default condition of internet access as the first restricted cohort reaches adulthood

Age assurance ceases to be a compliance feature of particular services and becomes a default precondition of consumer internet access in the largest…

Claude · 2031 · plausible

Prior state

Australia's prohibition on social media accounts for those under sixteen took effect in December 2025, followed across the second half of the decade by variants in the United Kingdom, several European member states, and a bloc of United States states, most of them carrying statutory review clauses and sunset or reporting obligations. In parallel, the European digital identity wallet obligations came into application, giving the largest regulated market a state-issued attribute-attestation mechanism suitable for age claims. Device-level versus platform-level assurance remained contested between operating system vendors and social platforms.

Material change

Age assurance ceases to be a compliance feature of particular services and becomes a default precondition of consumer internet access in the largest regulated markets: a verified or attested age signal is required at account creation across the majority of the most-used consumer services in those jurisdictions, and pseudonymous account creation becomes a minority pathway rather than the norm. The same rails are extended by statute or by commercial practice to gambling, adult content, alcohol and nicotine sales, AI companion services, and in some jurisdictions to general account provenance and content-authenticity labelling. This occurs irrespective of the outcome of the first evaluations, which is the point of the development.

Why now

Two clocks land together. The cohorts that were thirteen when the first restrictions took effect reach adulthood around 2030 and 2031, producing the first five-year longitudinal comparison between restricted and unrestricted populations of the same age — before this, the evaluation literature could only measure short-run displacement. Simultaneously, the statutory review and reporting obligations written into the first-wave instruments fall due in this window, requiring governments to publish findings and to decide on renewal, extension, or repeal within the year. The infrastructure question is settled in 2031 because that is when the political decision cannot be deferred, not because the technology matured on a schedule.

Mechanism and resistance

The mechanism is regulatory cost allocation: once a service must verify age for one population, verifying for all is cheaper than segmenting, and once a state issues attestation credentials, private verification becomes commercially irrational. Network effects then do the rest, because a user who has an attested credential for one service supplies it to others without friction. Resistance is substantial and partly successful. Constitutional litigation in the United States constrains the breadth of state mandates; data-protection authorities in Europe impose data-minimisation requirements that push assurance toward zero-knowledge attestation rather than document upload; circumvention through virtual private networks and unregulated services is easy for motivated adolescents and empirically significant; and small services cannot comply, so enforcement concentrates on large incumbents and drives usage toward the periphery. Where the evaluations are negative, repeal campaigns gain force — but they target the restriction, not the infrastructure, which by then serves other purposes.

Consequences

The winners are the incumbent platforms whose compliance capacity becomes a moat, the identity providers who intermediate attestation, and the device vendors who win the architectural fight. The losers are small and pseudonymous communities, users with irregular immigration status or without documents, and populations in states that adopt the same rails without the data-protection constraints that motivated them in Europe — for whom age assurance functions as general-purpose identification of internet users. Adolescent media consumption does not decrease so much as relocate, toward messaging, gaming, and conversational AI services that fall outside the statutory definitions, which is the finding most likely to complicate the evaluations. The durable historical significance is architectural: the internet's default of unverified access, which had survived thirty years of proposals to end it, ends as a side effect of child protection.

End state

A consumer internet in the largest regulated markets in which identity attestation at account creation is normal, the child-protection rationale that produced it is empirically contested, and the infrastructure is already in use for purposes unrelated to children.

Observable test

The share of the most-used consumer services in Australia, the European Union, the United Kingdom, and the largest United States states requiring an age signal at account creation, compared with 2026; the published statutory review reports on the first-wave restrictions; the issuance volume of state-backed age attestations; and adolescent mental-health service presentations in the restricted cohort against comparison jurisdictions.

Disconfirming sign

The statutory reviews find no measurable effect, the first-wave restrictions are repealed or left unenforced, and age assurance remains confined to gambling and adult content.

Themes

Society & culture, Law & institutions, Security & cyber