The UK Home Office is pushing ahead with plans to use AI technology to guess the age of young people arriving at UK borders to seek asylum, starting in 2027. Yet the Home Office’s own tests found the technology performed worse on certain groups of people, notably Africans. The plans severely endanger the human rights of children seeking asylum and should be scrapped.
Facial age estimation technology (FAE) is a nascent technology used to estimate a person’s age, which would contribute towards determining their asylum status. Described with much fanfare by the Home Office as a “cutting-edge AI tech,” FAE is currently used in UK shops and bars on customers seeking to buy age-restricted items. To use this for life-changing decisions in refugee processing centers is to introduce an unreliable, untested technology into an already flawed process.
Human Rights Watch, Foxglove, and 61 other civil society groups have written to the Home Office asking them to halt these plans immediately as they create new, unnecessary risks to vulnerable young people. This technology has no place in deciding whether a young person can access the rights and protections they are entitled to.
We are asking the Home Office to address urgent questions around accuracy, efficacy, and discriminatory risks of the system, as well as a lack of legal justification, adequate safeguards, and accountability mechanisms.
The Home Office’s troubling justification for using this technology wrongly pitches FAE as a magical solution to complex issues and as a way increase deportations, while painting asylum seekers as fraudulent. This does not give confidence that young people subjected to a FAE assessment will be given fair and balanced treatment.
The Home Office guidance says the technology will advise human decisions for now. However, too often pilot tests of harmful technology serve as a gateway to wider adoption, as with the Metropolitan Police's trial of facial recognition technology, expanded incrementally over a decade to current widespread always-on surveillance across extensive use cases. As such systems are entrenched and expanded, the risk of automation bias increases, and the technology’s role often changes from advisory to authoritative.
No other government appears to use FAE in this way, and there is a risk that this will set a global trend. The UK is influential in setting migration policy, for example, the now-shelved Rwanda Scheme, since mimicked by Dutch and US governments. The FAE plans are simply too risky to pursue, and those risks extend far beyond UK borders. We urge the UK government to reconsider.
