Concerns Arise Over AI Use in Age Assessment for Asylum Seekers
A coalition of over 100 organizations dedicated to the rights of refugee children is raising alarms about a controversial initiative to deploy artificial intelligence (AI) for estimating the age of young asylum seekers. The coalition warns that this technology may lead to more instances of children being mistakenly placed in adult prisons and detention facilities.
This warning follows the Home Office’s recent announcement of a partnership to implement AI facial age estimation technology aimed at young asylum seekers with disputed ages. Critics underscore the potential negative consequences of such a measure, particularly for vulnerable youth who may not conform to typical age benchmarks.
A report by the Consortium for Refugee and Migrant Children—set to be published in June and obtained by The Guardian—cautions against applying AI to young individuals whose physical appearance is influenced by factors like trauma, malnutrition, and the arduous journeys they often undergo to seek safety. The report, titled “Benchmarks and Boundaries: Using Facial Age Estimation to Assess the Age of Unaccompanied Young People Seeking Asylum,” does not outright reject the use of AI. However, it emphasizes that AI should complement, not replace, thorough age assessments conducted by qualified social workers.
Legal guidelines stipulate that the Home Office must employ AI technology in an advisory capacity rather than as a definitive tool, ensuring a range of protections are in place, such as access to appropriate adults, legal advice, and the right to contest decisions.
The complexities surrounding age assessment are significant, especially since most unaccompanied minors seeking asylum in the UK are aged 16 or 17. According to Home Office data, social workers are more than twice as likely to classify these young asylum seekers as children compared to immigration officers at the border, with over two-thirds assessed as minors. This disparity raises vital questions about the reliability of age evaluations.
The Home Office’s announcement emphasizes the necessity of deterring adults who misrepresent their age to exploit the system, while concurrently acknowledging the imperative to safeguard the welfare of actual minors. Alex Norris, Minister for Border Security and Asylum, remarked that adult migrants misrepresenting their ages have long misused the system, diverting critical support from at-risk children. The ministry claims that AI technology will help identify and expedite the removal of those abusing the asylum process.
Final determinations will rest with immigration officials, who will ensure that the AI system undergoes comprehensive testing and validation before national implementation. Kamena Doering, co-chair of the Consortium for Refugee and Migrant Children, expressed significant concerns regarding the government’s proposals, noting that AI cannot adequately consider the myriad factors affecting the appearance of youth fleeing conflict or persecution.
Research supports the assertion that AI can exhibit biases and inaccuracies akin to human assessment errors. Khama Petrchenko, a senior policy analyst at the Refugee Council and consortium member, highlighted troubling statistics indicating that current visual assessments at the border often result in children being inaccurately classified as adults, leading to severe repercussions for their safety and well-being. She warned that employing flawed technology could automate existing errors, putting even more children at risk of being unfairly categorized and placed in inappropriate accommodations.
The Home Office plans to enhance the AI’s capabilities, using an initial analysis of facial images captured when young individuals arrive in Dover. A contract worth £322,000 over three years has been awarded to Akhter Computers Ltd, with hopes that the technology will undergo further testing before its introduction in 2027.
