Human Rights Network blasts Home Office plan to use racist AI on asylum-seeking children

The Human Rights Network has added its voice to the growing number of advocacy organisations and watchdogs which have denounced Home Office plans to use AI facial-recognition software to assess the age of asylum-seeking children.

The charity has already worked with hundreds of minors whom the state has wrongfully catagorised as adults.  Back in April, the Human Rights Network exposed the fact that the Home Office frequently detains so-called “age-disputed children” as adults.

Of the 76 age-disputed detainees at the time, Social Services had reassessed (or were reassessing) 26 as children.

Less than a month later, the Helen Bamber Foundation revealed that the Home Office falsely identified 755 children as adults in 2025 alone.

Home Office choosing biased, racist tech

However, at the end of May, the Home Office announced plans that would further compound the problem by using AI facial-age estimation (FAE) in cases where an asylum seekers’ age is in dispute. The software was developed by German facial-recognition company Cognitec Systems.

At the time, the Refugee and Migrant Children’s Consortium warned that the scheme could lead to yet more wrongful detentions. Technology monitoring group Foxglove also urged the government not to test experimental software on children.

In June, Lighthouse Reports published data from a leaked Home Office report which demonstrated that even FAE tech from the “best performing supplier” is deeply unreliable. The software consistently predicted that 17-year-olds were 18. It also showed distinct bias against Sub-Saharan Africans, along with a bias against girls.

It was also unclear if Cognitec was in fact the “best performing supplier”. As such, the possibility remains that the Home Office chose to run with an even worse (but cheaper or more convenient) option than the junk tech it actually tested.

‘This is not about protecting children’

Most recently, among these many critics of FAE, Maddie Harris of the Humans for Rights Network stated that AI wound do nothing to improve the number of false identifications among asylum-seeking children.

In a searing critique, Harris accused the government of trying to justify its broken system:

This is not about protecting children – this is about shoring up the decisions that [the authorities] are making.

Harris went on to add that the UK’s immigration system already adultifies children from countries like Somalia and Sudan. She stated that:

As soon as a child is treated as an adult, they are susceptible to – or exposed, rather – to any of the egregious consequences that the Home Office is applying to adults, such as detention and removal.

By falsely labelling them as adults, the Home Office can place these vulnerable minors in greater danger by detaining them alongside adults and deporting them more easily.

’30-month margin of error’

Worse still, the Guardian reported that:

The government has admitted that even the best systems can have a 30-month margin of error.

As a reminder, the Home Office intends to use FAE specifically when a minor’s age is in dispute. Even a 6-month margin of error should be inexcusable for such borderline cases when a child’s life is on the line.

And yet, the government clearly considers plus or minus 2.5 years inconsequential, because it just does not value asylum-seekers’ lives.

This is the ‘acceptable’ face of white supremacy, pure and simple.

Featured image via the Canary

By Grace

Source link