UK to Use Facial Age Scans on Asylum Seekers, Tests Show Flaws

The British government announced plans in July 2025 to use facial age estimation to help determine asylum seekers’ ages, but the Home Office has since delayed rollout until 2027 and a leaked April 2025 internal test report shows the technology produces significant errors. The report tested seven algorithms against more than 2.5 million images and found the best-performing system had “substantial deviations,” including an average 4.6-year error for female Sub-Saharan Africans.

What did the Home Office tests find?

The leaked April 2025 report evaluated seven facial age estimation algorithms on a dataset of more than 2.5 million images and found demographic and image-quality weaknesses: the unnamed best-performing algorithm performed significantly worse on Sub-Saharan Africans, tended to predict that some 17-year-olds were over 18, and did worse on females, with female Sub-Saharan Africans off by an average of 4.6 years. The report also says most of its testing used high-quality, documented photos, and that initial-encounter photos from border crossings were “routinely worse” than follow-up images, which likely worsens real-world accuracy.

How will facial age estimation be used at the border?

The Home Office frames face scans as an additional tool to support border officers’ judgments at the “first encounter,” and its guidance says that in cases of uncertainty “individuals will always be treated as children until a further assessment is conducted.” The internal report, however, leaves operational details unresolved: it says how FAE would be used in an operational context was still being explored, and it highlights that stress and the “temporary aging” effects of travel trauma appeared to affect algorithmic estimates. The department spent more than $400,000 in May on face-scanning technology from German company Cognitec, one of the vendors included among the seven algorithms tested.

What are the other technical and institutional concerns?

National Institute of Standards and Technology testing and the Home Office’s own analysis both point to two recurring problems: algorithmic bias by demographic group and strong sensitivity to photo quality. The leaked report notes it could not determine whether poor photo quality or the physical condition of asylum seekers at arrival had the greater impact on errors. A former scientific adviser who had been on a Home Office committee described the face scans as “hideously inaccurate,” and rights groups led by Foxglove, joined by 61 organizations, have urged the government to abandon the plans.

Why it matters

Misclassifying children as adults removes legal protections and can place young people in adult-only detention settings. Since 2010, 40 percent of people who have faced age assessments have been classed as adults, a proportion that underscores how consequential initial decisions are. Deploying a system that tests worse on the largest cohorts arriving via small boats, and that overestimates the ages of some minors, amplifies harms for groups already vulnerable after dangerous journeys.

What to watch

Observe whether the Home Office publishes operational standards about photo quality and staff training, and whether it releases details about which algorithm it plans to deploy before the 2027 rollout. The next confirmed milestone will be any formal procurement or operational guidance that specifies where, when, and how face scans will be used at the point of first encounter.