The government is pressing ahead with an AI tool used to check the ages of migrant children despite warnings from its own advisers that it is “hideously inaccurate”, said The Independent.
The Home Office announced last month that facial age estimation (FAE) technology would be used by immigration officers from 2027 to “crack down on fake claims by small boat arrivals posing as children”, but two reports have identified serious problems with it.
How does FAE work?
AI-powered FAE tools have been used to prevent children from accessing age-restricted goods and services, including cigarettes, alcohol and adult-only online content.
The Week
Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.
SUBSCRIBE & SAVE
Sign up for The Week's Free Newsletters
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.
Latest Videos From
Now, the “hardening politics around asylum” has opened a new market with “far higher stakes” – using AI to determine whether undocumented migrants are under or over 18, said Lighthouse Reports, in collaboration with The Independent and Wired.
A photograph is fed into an AI system and goes through several layers of analysis, each picking up “increasingly subtle patterns”, said Oli Buckley, a cybersecurity professor at Loughborough University, on The Conversation.
The system, which is “trained on millions of photographs of people whose ages are already known”, has learned to associate patterns in a face with likely age ranges. It studies skin texture, the depth of lines around the eyes, bone structure and the distribution of soft tissue, before delivering a “probability distribution”, which is closer to “most likely between 17 and 21” than “this person is 18.”
All of this matters because at the UK border, deciding whether someone is 17 or 19 is a very “consequential judgment”. If you “get it wrong one way”, a “vulnerable child” loses legal protections they’re “entitled to”, but “if it’s wrong in the other direction”, then an adult “enters a system designed for minors”.
What are the issues?
The technology misclassified more than a third of 16-year-olds as adults and in some tests it was shown to give the wrong assessment in 70% of cases, according to an audit by Lighthouse Reports.
A separate, leaked report, which the Home Office tried to withhold, found the technology is least accurate when trying to assess migrants from countries such as Eritrea and Sudan, which have the highest number of small boat migrants arriving in the UK. This has led to accusations that the technology has “baked-in racial bias”, said The Independent.
The report also warned that “error rates are particularly high for female child migrants from Sub-Saharan Africa”. Estimations were out by 4.6 years on average – “meaning a 14-year-old girl could be predicted to be an adult”. The AI tool could also be less accurate for people with visible ageing caused by “stress of travel”.
Will it still be used?
The Home Office said it has “rigorous processes in place to verify an individual’s age” and the “groundbreaking assistive tool is designed as an additional source of information for immigration officers, and does not replace or overrule human judgment”.
But Anna Bacciarelli, senior researcher on technology, rights and investigations at Human Rights Watch, told Lighthouse Reports that this was of little reassurance because of the “well-established phenomenon” of automation bias, where people tend to trust a computer’s decision over their own judgment.
The UK’s use of the technology sets a “dangerous precedent”, she said, and “other countries at entry points across Europe are likely to follow suit”. As a result, “use of this inaccurate and invasive technology could become widespread”.
