Charities warn UK government's AI age-detection system for migrants risks misclassifying child refugees

Concerns Over AI Age-Detection Implementation

The British government intends to introduce AI-powered facial age-estimation technology to screen migrants. However, children's charities and rights groups are raising alarms, stating that the underlying models are flawed and racialized, posing a significant risk to child refugees.

Critics are urging ministers to reconsider these plans, specifically highlighting the danger that Black children arriving from conflict zones could be incorrectly categorized into the adult immigration system.

Existing Issues and AI's Potential Impact

Maddie Harris of the Humans for Rights Network noted that her charity has already encountered hundreds of children wrongly placed into the adult system through current age assessments by immigration officers. The Helen Bamber Foundation reported that 755 children were incorrectly identified as adults upon arrival in the UK in 2025, according to Home Office figures.

Harris believes that introducing AI is a "cheap process" that will not improve these numbers, suggesting it serves to reinforce existing decisions rather than protect children. The government itself acknowledges that even the most advanced systems can have a 30-month margin of error.

Risks for Vulnerable Children

The margin of error is particularly risky for children, especially those from countries like Sudan and Somalia, who are often "adultified" by the current system. If treated as adults, these children become susceptible to severe consequences such as detention and removal, which are typically applied to adults.

Unaccompanied children misclassified as adults could be forced into accommodation alongside adults, increasing their exposure to risks and potential violence, especially given the rise in protests targeting asylum seeker accommodations.

Technology Provider and Bias Concerns

In May, the government announced a contract with Cognitec, a German firm specializing in face recognition, to provide the facial age-estimation technology. An investigation by Lighthouse Reports in June indicated that this system exhibits bias against sub-Saharan Africans, tending to overpredict the ages of minors.

UK-based tech monitor Foxglove has criticized these systems for relying on "racist bias" and has advised against using children to test experimental technology. The Refugee and Migrant Children’s Consortium has also expressed concerns about the technology's ability to accurately assess children who have experienced difficult journeys, noting that trauma can affect a child's appearance and potentially lead to misjudgment.