UK AI age‑detection tool risks misclassifying child refugees as adults, raising bias and legal concerns
Executive summary: The UK government plans to roll out an AI‑powered facial‑recognition system to assess the age of asylum‑seeking children, but independent tests show the model systematically overestimates age for non‑white faces. Misclassifying minors as adults can result in children being housed with adults, exposing them to risk and potentially violating UK and international child‑protection laws.
Who is involved: UK Home Office, Human Rights Network charity, AI vendors supplying the facial‑recognition software, and refugee advocacy groups.
Likely next: Parliamentary committees may request an impact assessment; the Home Office could pause the rollout pending bias audits, and regulators may examine the tool under the EU AI Act’s high‑risk provisions.
The Guardian reports that facial‑recognition software used by the UK Home Office to estimate age shows a racist bias that overestimates the ages of child refugees, potentially placing them in adult accommodation. Human Rights Network warns this flaw could lead to safeguarding breaches and legal challenges under equality legislation. The story highlights growing scrutiny of AI systems deployed in public‑sector decision‑making.
Analysis — what this means
Likely next events
- UK Home Office to publish bias audit results by 15 August 2026
- Parliamentary Joint Committee on Human Rights to hold hearing on AI age‑detection in September 2026
Sectors affected
- Public sector immigration and asylum services
- AI facial‑recognition technology vendors
Regulatory implications
- Potential application of the UK Equality Act 2010 to challenge discriminatory AI outcomes
- EU AI Act may classify the system as high‑risk, requiring conformity assessment
Historical parallels
- 2020 UK Home Office visa‑allocation algorithm bias controversy that led to policy revision