Part E of the Age Assurance Technology Trial focuses specifically on age inference – a method of determining an individual’s likely age or age range based on verifiable contextual, behavioural, transactional or environmental signals, rather than biometric data or identity documents. Unlike age verification, which relies on a known and validated date of birth or age estimation, which uses biometric characteristics to predict age, age inference draws reasonable conclusions about age by analysing facts such as school enrolment, financial transactions, content barring settings, service usage or participation in age-specific activities.
Age inference is viable and effective for age assurance in Australia, especially when used to flag likely underage access, support early safety interventions or trigger fallback mechanisms. When designed with transparent logic, strong input signals and bounded confidence thresholds, inference systems offer low-friction, proportionate assurance.
There are no substantial technological limitations to deploying age inference systems. Many providers demonstrated the ability to operate on existing platform data (e.g. interaction logs, metadata, using mature inference engines or rule-based models). System effectiveness depends not on technology maturity but on the relevance, diversity and quality of input signals.
Inference methods were most accurate when grounded in clearly modelled reasoning and when drawing from well-labelled behavioural signals. For example, systems using language complexity, session timing or feature access patterns were able to classify age thresholds (e.g. under 13, under 18) with high reliability in controlled and real-world scenarios.
Age inference is inherently context specific. It must be tailored to the sector, risk profile and digital behaviours of the user group. Inference approaches were most successful when adapted to the norms of child-facing platforms, regulated services or role-specific environments (e.g. education, gaming, retail).
The age inference sector is innovative and rapidly evolving, with providers exploring techniques like gesture modelling, narrative complexity analysis and contextual metadata synthesis. Some systems demonstrated promising accuracy using only temporary, non-identifying inputs, supporting the shift toward Zero-Knowledge assurance.
Security and governance of inference systems were generally strong, particularly among independent providers and those using in-session logic with no persistent data. However, platforms deploying inference as part of ongoing behavioural monitoring must address the risk of profiling, surveillance or footprint expansion, as flagged in ISO/IEC FDIS 27566-1.
Inference quality depends on the transparency and reasonableness of the underlying logic. Systems that disclosed their signal-to-inference mapping, applied confidence thresholds conservatively and included fallbacks or escalation paths performed more ethically and effectively, with fewer classification errors.
Fairness and demographic sensitivity remain active areas for improvement. Some systems risked bias where signals correlated with culturaly linguistic or socioeconomic factors. Providers are encouraged to conduct demographic impact assessments, improve calibration and align with inclusion clauses of ISO/IEC FDIS 27566-1 to mitigate false
positives.
Verifymy provides flexible AV solutions integrated with digital wallets, document verification and cross-jurisdictional datasets. It supports selective disclosure and privacy-first age checks, delivering binary outcomes (e.g., “Over 18: Yes”) via APIs and reusable credentials for platforms such as gambling, e-commerce and education.
Aggregates age assurance services, enabling inference through third-party integrations; supports orchestration logic for fallback, confidence thresholds and multi-vendor identity decisioning.
Luciditi provides facial age estimation, document verification via selfie-ID match, NFC passport reading and open banking or telco records, with fallback to a reusable digital ID app.
Yoti provides low-friction, high-trust verification with one-time & reusable tokens. A standout example of minimising user friction while maintaining assurance comes from Yoti, whose platform consistently prioritised privacy, simplicity and user control throughout the Trial.
PRIVO provides privacy-focused, parental consent-based AV services using facial estimation, document checks and guardian approval workflows. COPPA-certified and focused on protecting children in online services and educational contexts.
Erratum: The published report suggests the solution is “US-Centric” – while it has been designed specifically to cater to the US market, and the COPPA legislation in particular. We acknowledge that PRIVO’s solution can be configured for other jurisdictions, including Australia
Erratum: The MyMahi solution was included in the Age Inference volume of the published report, misunderstanding that their assessment of age is based on school year group. MyMahi verifies age from the date of birth held in school records.
