Starting next year, the UK government plans to use AI-powered facial age estimation on asylum seekers arriving at the border. It’s believed to be the first time this technology has been deployed for immigration purposes anywhere.
The stakes are enormous. Many asylum seekers don’t have documents proving their age. If the AI classifies a child as an adult, that person loses legal protections and can be placed in adult-only detention centers.
An investigation by WIRED and Lighthouse Reports, working with The Independent, obtained an internal UK government report detailing tests of facial age estimation systems. The results aren’t reassuring. The systems regularly mistake children for adults and show serious bias problems — particularly affecting the largest group of migrants subject to age assessments in 2025.
This isn’t happening in a vacuum. Age verification is already spreading across the internet — social media bans in Australia, porn restrictions in half of US states. But applying it at a border, with life-changing consequences for vulnerable people, is a different proposition entirely.
The British government knows the internal test results. It’s proceeding anyway.
Facial age estimation has been criticized by researchers for accuracy issues and demographic biases. Deploying it in a high-stakes immigration context, where errors directly harm children, raises questions that go beyond technical performance. It’s about whether this technology should be making decisions that affect whether someone is treated as a minor or an adult.
The Home Office data shows which migrant groups are most affected. The bias in the testing directly impacts them. That’s not a minor footnote — it’s the core problem.
