UK Home Office Advances Facial Age Estimation for Asylum Age Assessments

The UK Home Office has published guidance on facial age estimation for immigration age assessments, setting out how the technology could be used to help determine whether an undocumented arrival should be treated as a child or an adult. The guidance states that facial age estimation tools are not currently used in the immigration system and that the Home Office is exploring how they could strengthen the age-assessment process, describing how the technology works along with its benefits and limitations.

Age assessment is one of the more contested decisions in the asylum system. When an arrival has no reliable documentation, immigration officers must make an initial judgment about whether the person is under 18, a determination that governs where they are housed, how their claim is handled, and what safeguarding protections apply. Facial age estimation uses an algorithm trained on large image sets to estimate a person’s age from a photograph of their face, producing a numerical estimate rather than a yes-or-no identity match.

Alongside the guidance, the Home Office has moved the work toward operational testing through a procurement. A three-year contract worth 322,000 pounds, beginning June 1, names British IT supplier Akhter Computers as the main contractor, with the facial age estimation technology to be supplied by German biometrics firm Cognitec as subcontractor. Reporting on the award indicates a live trial is planned at the Western Jet Foil facility in Dover, with a wider rollout targeted for mid-2027, contingent on whether the published accuracy and error rates can withstand legal and ethical challenge.

The plan has drawn criticism from rights and professional groups. Human Rights Watch has called the scheme cruel and unconscionable, arguing that facial age estimation is unproven in refugee settings, and the British Association of Social Workers has warned that introducing the technology into age assessment could lead to safeguarding mistakes. The accuracy of age-estimation algorithms varies with the age range being tested and the demographics of the person in front of the camera, which is the core of the concern when the output feeds a decision about whether someone is a child.

Facial age estimation is already being trialed elsewhere in the UK in lower-stakes settings, including retail self-checkout age checks for age-restricted purchases. The asylum use case raises the stakes considerably, applying the same class of technology to a decision with direct consequences for a vulnerable population.

The Home Office guidance frames the technology as one input to the age-assessment process rather than a sole determinant.

Sources: UK Home Office, The Independent, LBC

–

By Ali Nassar-Smith