Retroactive Age Verification: The Existing-Account Problem

Retroactive age verification is the requirement to establish the age of users who already hold accounts, rather than only checking new registrations. From 1 September 2026 Brazil bars the creation of new social media accounts for users under 15. From 1 January 2027, existing accounts belonging to under-15s must be age-verified or deactivated.

Every age assurance regime so far has been argued over the first requirement. The second is the one that breaks operating models.

Key takeaways

• Brazil's phased rule creates a four-month window between the new-account ban and the existing-account deadline.
• Checking new signups is a flow problem: verify at the gate, and the population self-selects into compliance over time.
• Checking existing users is a stock problem: an entire installed base must be assessed at once, including people who joined years ago under different terms.
• Deactivation is the default consequence, which makes a false negative commercially expensive and a false positive reputationally expensive.
• The economics only work if a user can prove their age once and reuse that proof, rather than every platform running its own capture against its whole base.

Flow versus stock

The distinction determines almost everything about implementation cost.

At the gate, a verification step sits inside a moment where the user is already motivated to complete something. Applied retroactively, the same step arrives unprompted, interrupts a service the person already uses, and asks for sensitive information with deactivation as the stated alternative. Completion rates are not comparable, and the support burden falls in a compressed window.

Three ways platforms will try to do this, and what each costs

Blanket re-verification. Ask everyone. Simple, defensible, and the most expensive option in both direct cost and churn. It also creates the largest possible pool of newly collected identity data, at exactly the moment regulators are scrutinising retention.

Risk-based targeting. Use behavioural and account signals to estimate which accounts are likely to be under-age, and verify that subset. Cheaper and less intrusive, but inference about minors is itself sensitive, and an inaccurate model produces both missed accounts and wrongly challenged adults.

Credential acceptance. Accept a reusable, privacy-preserving age proof the user already holds. Lowest friction where such credentials exist, and it collapses the cost for the user across every platform asking the same question in the same window.

The third option is the only one that improves as more regulators adopt similar rules. The first two get worse: each new jurisdiction adds another full-base campaign.

Why the data-minimisation constraint is not optional here

Retroactive verification collides with a second obligation. A platform running a full-base age check is collecting identity documents from a population that includes a large number of minors, and storing the result creates precisely the kind of artefact archive whose consequences we examined in the case against storing documents.

A defensible approach establishes the age band and retains only the outcome, not the evidence. Zero-knowledge age proofs make this concrete: the platform learns that the user is over or under a threshold, and learns nothing else. No date of birth, no document, no image.

AIR Identity supports exactly this pattern. A user verifies once with a trusted issuer and holds a reusable credential. Any platform asking the age question receives a cryptographic proof of the threshold it cares about. The user is not re-verified per platform, and no platform accumulates a document store. For communities and fandom platforms with large existing user bases, that difference determines whether a retroactive requirement is a project or a crisis.

The accuracy dimension matters alongside it: as we covered in age assurance accuracy standards, regulators are increasingly specifying error tolerances rather than accepting best effort.

Preparing for the stock problem

Segment the base before the deadline, not during it. Establish now which accounts have a reliable age signal, which have a weak one, and which have none. The third group is your actual workload.

Decide the deactivation policy in advance. What happens to an account that does not respond? To content and social graph? To a user who is wrongly flagged and appeals? These become support and press problems if decided reactively.

Instrument the funnel. Retroactive verification has a completion rate, and it will be lower than signup verification. Knowing it early determines whether the campaign needs to start in month one or month three.

Accept external credentials where you can. Every user who arrives with a valid proof is one you do not have to verify, support, or store data about.

Watch the other jurisdictions. Brazil is not an outlier. Comparable regimes are live or advancing across Australia, the European Union, the United Kingdom and multiple US states, as covered in under-16 social media age verification. Building a single-jurisdiction solution guarantees rebuilding it.

Frequently asked questions

What does Brazil's social media age law require?

From 1 September 2026, no new social media accounts may be created for users under 15. From 1 January 2027, existing accounts belonging to users under that age must be age-verified or deactivated.

What is retroactive age verification?

It is the requirement to establish the age of users who already hold accounts, rather than only verifying new registrations. It differs from signup verification because the entire existing user base must be assessed within a fixed period, and users have no immediate incentive to complete the check.

Why is verifying existing accounts harder than verifying new ones?

New users complete verification as part of gaining access, so motivation is high and the cost is spread across time. Existing users are interrupted in a service they already use, motivation is low, completion rates are lower, and the whole base must be processed before a statutory deadline.

How can a platform verify age without collecting identity documents?

By accepting a zero-knowledge age proof from a credential the user already holds. The platform receives confirmation that the user is above or below the relevant threshold and receives no date of birth, document or image, so nothing sensitive is stored.

Do age verification rules apply outside the country that passed them?

Frequently, yes. Most of these regimes attach to services offered to users located in the jurisdiction, regardless of where the platform is established. A platform with users in multiple regulated markets is typically subject to several regimes at once.

Related reading

• Under-16 social media age verification
• Age assurance accuracy standards in 2026
• The ID scan breach and the case against storing documents

More from AIR: AIR Identity, communities and fandom, or browse the full AIR blog.

Facing a retroactive age requirement across an existing user base? See how AIR Identity verifies an age threshold without collecting documents, or talk to our team.