A comprehensive government-commissioned study on online age verification has revealed that while effective methods exist, a universal solution does not. The research highlighted significant concerns about privacy, potential biases, and accuracy, dependent on the verification approach used.
According to the findings, age verification can be performed “privately, efficiently, and effectively,” but cybersecurity issues persist, along with unnecessary data retention by some providers. Communications Minister Annika Wells noted that the report demonstrates the technology’s capacity to better protect young users from harmful content. She emphasised that while a one-size-fits-all solution is unattainable, various effective options exist that can prioritise user privacy.
The report suggests that age verification methods, particularly those checking against government documents, are generally reliable, bolstered by strong security protocols. Some technologies successfully detected AI-generated fakes, although not all service providers could perform live checks with government databases.
However, there are worries that certain providers are over-retaining user data, including biometric specifics, without justification. This approach raises significant privacy concerns and necessitates clearer regulatory measures. Alternatively, age estimation methods that evaluate a user’s appearance or behaviour—rather than requiring sensitive documents—could mitigate such risks.
The report, produced by the UK-based Age Check Certification Scheme (ACCS), described age estimation as a “frictionless, privacy-conscious” strategy. Nevertheless, it noted that even the best systems only achieve an average accuracy of 1.3 to 1.5 years, leading to uncertainty about regulating access strictly based on age.
The authors cautioned that it’s a misunderstanding of age estimation technology to expect precise age-restrictions without accepting potential margins of error, necessitating alternative methods for confirmation. The technology may also struggle with effectively identifying users from non-white demographics.
Despite the challenges, the study explored various technologies from 48 providers, considering age verification, estimation, and inference, alongside approaches that combine multiple methods. However, it clarified that it does not endorse specific policy recommendations or technologies.
As part of an upcoming social media ban aimed at protecting children under 16, platforms like Facebook, Instagram, TikTok, and others must implement appropriate measures by December 10. Non-compliance could lead to hefty fines of up to $49.5 million. This study serves as an important step toward enhanced online safety for younger users while highlighting the complexities involved in age verification.