UK Home Office AI age-estimation system risks misclassifying refugee children as adults
The UK Home Office announced plans in May 2026 to use Cognitec's AI facial age-estimation system to screen asylum-seeking migrants at borders. Rights groups and charities warn the technology exhibits racial bias against sub-Saharan African children, reportedly overestimating their ages and risking their misclassification as adults, which would expose them to detention, removal and unsafe accommodation conditions.
What happened
The UK Home Office announced in May 2026 a contract with Cognitec, a German facial-recognition firm, to introduce AI-powered age-estimation technology at borders to screen asylum-seeking migrants.
An investigation by Lighthouse Reports in June 2026 found the system showed bias against sub-Saharan Africans, tending to overpredict the ages of minors. The government has admitted that even the best systems can have a 30-month margin of error.
Reported impact
According to Home Office figures cited by the Helen Bamber Foundation, 755 children in 2025 were incorrectly identified as adults on arrival to the UK. Maddie Harris of the Humans for Rights Network stated that unaccompanied children would be forced into accommodation alongside adults, exposing them to risk and to violence.
Harris said that as soon as a child is treated as an adult, they are susceptible to detention and removal by the Home Office. She said children from countries such as Sudan and Somalia were already "adultified" by the system.
Index analysis
Foxglove stated that these systems relied on "racist bias". The Refugee and Migrant Children's Consortium raised concerns about the technology's ability to assess children, noting that stress from travel can age a child's face and invalidate the assumptions on which the technology is based.
Maddie Harris described the introduction of AI as a "cheap process" that would not improve the numbers, stating "this is not about protecting children – this is about shoring up the decisions that [the authorities] are making."
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.