UK: Home Office Should Halt AI Experiment on Children Seeking Asylum

Human Rights Watch joins 61 civil society groups in calling on the UK Home Office to halt its plans to use facial age estimation (FAE) technology to evaluate whether people seeking asylum are under or over 18. This experimental technology poses significant risks to the rights of children and all asylum seekers. In a letter to the Home Office, the groups outline concerns regarding discrimination, inaccuracy and inefficacy, potential privacy violations, and lack of transparency around the plans. The scheme was first announced in July 2025, and the Home Office expects to launch it in 2027. 

This open letter has been prepared by Foxglove and Human Rights Watch. A full list of signatories can be found at the bottom of the page. 

The Hon. Alex Norris, Minister of State for Border Security and Asylum  

CC: The Rt. Hon. Shabana Mahmood, Secretary of State for the Home Department 

2 Marsham St

SW1P 4DF 

Dear Minister, 

We write as 62 organisations alarmed at the Home Office’s decision to deploy Facial Age Estimation (FAE) to assess asylum seeking children from 2027. We believe this technology poses a serious threat to the rights of these children, a uniquely vulnerable population, and call on the Home Office to halt its use immediately.  

Discrimination: 

There are substantial and well-founded concerns about the bias of FAE. Many have highlighted its baked-in failures and discrimination, particularly in relation to women and people of colour. The Home Office has itself noted “that FAE performance can vary depending on ethnicity” and skin tone. 

Evidence shows FAE is most accurate when estimating the ages of Eastern European men but even then it consistently produces errors. On what basis can the Home Office possibly conclude this technology will assist in their decision-making for asylum seeking children who are predominantly people of colour? Please explain. 

Inaccuracy: 

The Home Office admits FAE systems are imprecise at the crucial 16-to-18-year-old boundary, with even the “top systems” having an “error margin of around 2.5 years here”. This is exactly the age range for which the Home Office has chosen to deploy this technology. And this error margin will be widened yet further as a result of the stress and trauma induced ageing from which asylum seeking children suffer. 

The faces of children seeking asylum tend to prematurely age from trauma and violence as well as months or years of travel enduring malnutrition, dehydration, sleep deprivation, and exposure to salt water during a dangerous sea crossing. Their facial features are likely to be distinct from other children their age who have not experienced trauma and violence. 

As such, unless this technology has been trained on asylum seeking children in particular – which itself raises questions about its lawfulness (please see below) – we can see no basis upon which the Home Office has concluded this technology will increase the accuracy of its decision making. Please explain. 

Lawfulness of Use of Children’s Data: 

Please confirm on what images and/or data this technology has been trained. We can see no lawful basis upon which the Home Office, or its chosen third-party FAE vendors, could have sought consent for the collection and processing of photographs or data from asylum-seeking children to train this system. If you disagree, please explain. And, because of the unique facial features of asylum seeking children, without the use of such unlawful data, we can see no basis on which the system could hope to produce accurate results. 

And therefore, we cannot see how the use of this technology to assess the age of asylum seeking children could ever be lawful. If you disagree, please explain on what basis.

Lack of Necessary Disclosure: 

We appreciate the publication of this guidance, however the lack of detail means we are unable to properly understand and/or assess how the rights of children and others seeking asylum in the UK will be safeguarded when FAE is in use. Please set out, in detail, how this FAE technology has been built and the data upon which it has been, and continues to be, trained.  

The Home Office claims “extensive testing has already been carried out across diverse groups, including different ethnicities, genders and age ranges, indicating promising performance and accuracy.” But these “promising” results have not been published. Why? The Home Office has also refused or failed to publish any Equality or Data Protection Impact Assessments. We write urgently seeking that information and to request the Home Office halt the scheme in the meantime.  

It is particularly concerning that the Home Office’s communications about the roll-out of FAE lean heavily on the argument that adults are masquerading as children to gain favourable treatment, when recent research suggests that children are currently being misidentified as adults and, in some cases, deported under the one-in-one-out policy. 

Please confirm:  

1. Has this AI system shown “promising” results for people seeking asylum whose facial ageing can present differently due to trauma suffered? Please provide us with a copy of these results, disaggregated by age, gender, country of origin, and any other relevant demographic information.
2. Please provide us with details about how this FAE was tested? Please set out the methodology used along with all Equality and Data Protection Impact Assessments completed to date. 
3. If the system has been trained and tested on people seeking asylum, on what basis was this data collected? From where? How was it processed? And with what consent process in place? 
4. At what stage(s) of the asylum process will this technology be deployed? What decisions will be made as a result of any age estimation created by this FAE technology?
5. How will the results be taken into account as part of any officers’ decision-making process?  
6. What opportunities will people seeking asylum have to scrutinise the results of FAE? What will the process be for contesting the results, and appealing any of the resulting decisions?  
7. What are the safeguards and protections you have in place to ensure this AI system does not infringe upon the rights of children seeking asylum? 

The Independent Chief Inspector of Borders and Immigration, in his 2025 report on the Home Office’s use of age assessments, was damning. He wrote, “I have listened to young people who felt disbelieved and dismissed by the Home Office, whose hopes have been crushed, and whose mental health has suffered.” The Home Office, he said, had an “uphill task” in persuading the British people that it could be trusted on age assessment. 

We ask now that the Home Office take on this uphill task in good faith and release the information that is required to scrutinise the legality and ethics of this deployment of new technology on some of the most vulnerable people – and children – in our world. 

We respectfully request a response from your department within 21 days.  

Yours sincerely,