TikTok will begin rolling out new age-detection technology across Europe in the coming weeks, the company said, as the ByteDance-owned platform faces mounting regulatory pressure to better identify and remove accounts belonging to children under 13.
The previously unreported system follows a year-long pilot in Europe and is designed to strengthen enforcement of TikTok’s minimum-age rules without resorting to blanket, intrusive checks. The company said the tool analyses profile information, posted videos and behavioral signals to predict whether an account may be underage. Accounts flagged by the technology will be reviewed by specialist moderators rather than automatically banned.
- Advertisement -
The rollout comes as European authorities intensify scrutiny of how social platforms verify users’ ages under strict data-protection rules, amid concerns that existing approaches are either ineffective or overly invasive. TikTok said the new technology was built specifically for Europe to comply with the region’s regulatory requirements and that it worked with Ireland’s Data Protection Commission, its lead EU privacy regulator, while developing the system.
Governments are moving quickly to set firmer boundaries on social media access. Australia has imposed what it calls the world’s first social media ban for children under 16, and Denmark has proposed banning social media for those under 15. In the United Kingdom, a pilot on TikTok led to the removal of thousands of additional accounts identified as belonging to users under 13, underscoring both the scale of the problem and the limits of traditional self-declaration methods.
Despite extensive efforts across the tech industry and by regulators, TikTok said there is no globally agreed way to confirm a person’s age while fully preserving privacy. The company intends to use multiple methods to handle edge cases and appeals. For users contesting an underage ban, TikTok said it will employ facial-age estimation from verification provider Yoti, along with credit-card checks and government-issued identification.
Yoti’s tools are also used by other platforms; Meta, for example, relies on Yoti to verify users’ ages on Facebook. These third-party services have emerged as a middle path between invasive identity checks and ineffective self-reporting, though they, too, face scrutiny over accuracy and safeguards for children’s data.
TikTok emphasized that its predictive system will not be used to automatically terminate accounts. Instead, it aims to help moderators triage risk signals more efficiently and act on likely underage accounts. The company said European users will be notified as the technology launches and expands across the region in the coming weeks.
The move highlights a broader shift in content moderation toward proactive detection, supported by machine-learning systems tuned to platform behavior. For TikTok, which has hundreds of millions of users and a heavy youth audience, better underage detection is both a compliance imperative and a reputational test. Regulators have warned that platforms must show measurable improvements in keeping children off services that are not designed for them, while parents and educators have called for clearer, less intrusive verification paths.
How well TikTok’s system performs—and how it balances effectiveness with privacy—will be closely watched by policymakers and rival platforms. With no universal standard for age assurance, companies are experimenting with a mix of signals, third-party estimators and ID checks, often tailored to regional legal frameworks. TikTok’s Europe-specific build reflects that patchwork reality, and its collaboration with Ireland’s privacy watchdog suggests ongoing regulatory oversight as the tool scales.
For now, TikTok’s message is incremental: more signals, more human review, and more regional tailoring. Whether that will be enough to satisfy Europe’s data-protection authorities—and keep under-13s off the platform—will become clearer as the rollout proceeds.
By Abdiwahab Ahmed
Axadle Times international–Monitoring.
