Online services are increasingly adopting age estimation technologies, which can determine a user's age using a selfie or phone number. Leading age assurance providers assert that these methods protect privacy since they do not link identity to estimated age or rely on existing data. Policymakers view this as a way to ascertain users' ages while maintaining their anonymity and safety.
Despite the absence of legal mandates, companies are voluntarily implementing these methods. For instance, Google plans to use machine learning to estimate ages by analyzing online activities, such as browsing history and YouTube views. Apple provides tools for parents to input children's ages when setting up iCloud accounts and restrict app access in the App Store.
Age estimation is considered an educated guess with potential errors. Age assurance providers often overlook what occurs when mistakes happen, potentially leading to secondary verification requirements.
Age assurance encompasses various techniques used by online service providers to determine user age. These include:
- Age declaration: Users provide their birthdate or affirm they are over 18.
- Age verification: More rigorous methods requiring ID scans or third-party database checks.
- Age estimation: Machine learning infers age based on characteristics like facial features or behavior.
Platforms may choose estimation methods like analyzing user data or facial age estimation, as Instagram does in partnership with Yoti through video selfies. However, when systems fail, users might need to submit sensitive information for verification, such as government IDs on Instagram or other identification documents on Roblox.
These predictive techniques often struggle with teenagers due to broad age ranges provided rather than specific ages. This poses challenges for enforcing legal access based on exact ages like 16 or 18 years old. Teenagers lacking government IDs may face difficulties accessing services without parental involvement.
Certain adult groups also face challenges when estimation methods fail. Nonbinary and trans individuals may be misclassified by facial technologies and lack appropriate IDs reflecting their gender and name. People with disabilities affecting appearance might be misclassified due to technological limitations. Facial recognition errors disproportionately affect people of color, leading to distrust of these systems. Additionally, low-income individuals and some immigrant communities may lack necessary documentation altogether.
These issues impact significant user groups globally who might be asked for additional personal data if misclassified by algorithms. Misclassification could lead adults interested in roleplay game reviews on YouTube being identified as children needing further verification steps involving sensitive documents.
A study following Louisiana's implementation of an age verification law showed users migrating towards services without such requirements—highlighting preferences against sharing IDs online while accessing benefits from digital platforms remains crucial.
Companies must prioritize protecting human rights across all demographics before deploying new assurance measures impacting privacy rights alongside essential online access needs worldwide.