Age assurance is the umbrella term for the ways an online service works out how old someone is. It spans age verification, age estimation and age inference, running from a full identity check down to an educated guess. In 2026 the gap between those methods carries legal weight, because regulators in the European Union, the United Kingdom and Australia now require services to gauge, with different degrees of confidence, whether a user is a child.

What do verification, estimation and inference actually mean?

The three methods trade accuracy against privacy and friction. Age verification settles someone’s age through a full identity process, usually against an official document or a trusted record, so it gives high certainty and collects the most data. Age estimation uses artificial intelligence to predict an age or age range with lower confidence, so a facial-age scan returns a band rather than a confirmed birth date. Age inference reads age from other signals, such as account history or a device setting, and offers the least certainty of the three. Pick the strictest check for every feature and you can end up excluding the young people the rules meant to protect, so the method has to match the risk.

What is the EU doing about age verification in 2026?

The European Commission has built a privacy-preserving age verification app and, on 29 April 2026, urged member states to make it available by the end of 2026. The app lets users prove they meet a required age threshold without revealing their exact age, identity or any other personal details, and a country can run it as a standalone app or fold it into a European Digital Identity Wallet. It supports the Digital Services Act, under which platforms must ensure a high level of privacy, security and safety for minors online. The design goal is to prove the age, not the person.

How strict are the UK’s age-check rules now?

Since 25 July 2025, services that show pornography or other content harmful to children have had to use highly effective age assurance under the Online Safety Act. Ofcom tests a method on four criteria: technically accurate, robust, reliable and fair. Approaches that can clear that bar include facial age estimation, photo-ID matching, open banking, credit-card checks and mobile-network operator checks, while a self-declared tick box does not. A service that fails can face fines up to 18 million pounds or 10% of global turnover. Early data reads both ways. Ofcom’s 2026 review reported that the share of children who met a highly effective age check rose from 25% to 43% between July 2025 and January 2026, while the regulator also raised doubts about whether current age-inference models could support a broader social media restriction.

What changes for under-16s in Australia?

Australia has gone furthest. Under the Online Safety Amendment (Social Media Minimum Age) Act 2024, age-restricted platforms have had to take reasonable steps to stop under-16s from creating or keeping accounts since 10 December 2025. eSafety treats Facebook, Instagram, Snapchat, Threads, TikTok, Twitch, X, YouTube, Kick and Reddit as age-restricted. Systemic non-compliance can cost a platform up to 49.5 million Australian dollars, and neither children nor their parents are penalised for getting around the rules. By mid-January 2026, platforms had deactivated, removed or restricted about 4.7 million accounts judged to be held by under-16s.

Does any single method work everywhere?

No single method clears every rulebook at once. A facial-age scan that satisfies Ofcom is not automatically the reasonable step Australia expects, and the EU app is not a national ban. The three regimes share a direction, which is safe and age-appropriate access rather than mass surveillance. A service reaches it by choosing the lightest method that meets each market’s duty, protecting whatever data it collects, favouring proof of an age band over a stored passport, and explaining the whole thing to parents in plain language.

Sources