Some assessments will inevitably be wrong, which is clearly a cause for concern
Beginning in mid-2027, the United Kingdom will trial artificial intelligence facial recognition at its Dover border processing centre to estimate the ages of asylum seekers, seeking to identify adults who claim to be children in order to access more favourable legal pathways. The technology, contracted for £322,000 over three years, enters a space already marked by imperfect human judgement — where current assessments find nearly half of child claimants to be adults, yet independent inspectors acknowledge that no method is without error. In the gap between catching deception and protecting the genuinely vulnerable, a profound moral question takes shape: what is the acceptable cost of being wrong about a child?
- With over 111,000 asylum claims filed in a single year and nearly half of child claimants found to be adults, the UK government is under intense pressure to close what it sees as a systemic loophole in its border system.
- The stakes of misclassification are severe — a child wrongly identified as an adult loses the legal protections and care system access the law guarantees them, a harm that independent inspectors say is not merely possible but mathematically inevitable.
- Human Rights Watch is calling for the scheme to be scrapped entirely, arguing the technology has only ever been tested in retail settings and has no proven track record in refugee processing environments.
- The Home Office has conducted preliminary ethnicity- and gender-diverse image testing but has not yet applied results to real decisions, positioning the Dover trial as a cautious first step rather than a full deployment.
- The technology is framed as a supplementary tool to assist officers rather than replace them, but critics warn that in practice, algorithmic outputs carry an authority that quietly displaces human judgement.
Starting in mid-2027, the UK Home Office will begin trialling an AI system that estimates asylum seekers' ages from photographs, developed by Harlow-based Akhter Computers Ltd under a £322,000 contract. The government's aim is to identify adults who falsely claim to be children — a misrepresentation officials say diverts resources from genuinely vulnerable young people and opens easier routes to remaining in the country.
The scale of the problem, by the government's own figures, is significant. Of more than 6,400 people who claimed to be children at the border in the year ending March 2026, 43 percent were assessed as adults. Minister for Border Security Alex Norris has framed the technology as a matter of fairness — catching those who game the system while protecting those who truly need it.
But an independent immigration inspector's report complicates that framing. Even trained officers using established methods — document checks, X-rays, MRI scans — sometimes get it wrong. The report was explicit: with no foolproof test available, some assessments will inevitably fail. When a child is misclassified as an adult, the consequences are not administrative but deeply personal — they lose the legal rights and protections the law was designed to give them.
Human Rights Watch has called on the government to abandon the scheme, with senior AI researcher Anna Bacciarelli describing it as "deeply flawed" and "cruel and unconscionable." The organisation notes the technology has only been tested in commercial settings like shops and bars — never in refugee processing contexts where the populations, pressures, and stakes are entirely different.
The Home Office says it has run preliminary tests on images representing the ethnicities and genders typical of asylum seekers, though those results have not yet informed any real decisions. The first live trial will take place at Western Jet Foil in Dover, with the AI functioning as a supplementary tool for officers when age is genuinely uncertain.
The deeper question the programme leaves unanswered is whether a system built to catch deception will, in the process, also misclassify children — and whether the promise of identifying more fraudsters can justify that risk.
Starting next year, the UK Home Office will begin testing an artificial intelligence system designed to estimate the age of asylum seekers by analyzing their photographs. The technology, developed by Harlow-based IT company Akhter Computers Ltd under a £322,000 contract, represents the government's latest attempt to address what officials describe as a persistent problem: adult migrants falsely claiming to be children in order to access the care system rather than the asylum system—a pathway that can make it easier to remain in the country.
The scale of the issue, according to Home Office data, is substantial. In the year ending March 2026, more than 6,400 people claiming to be children underwent age assessment at the border. Of those, 43 percent were found to be adults. The government argues that these false claims divert resources away from genuinely vulnerable young people. Minister for Border Security and Asylum Alex Norris framed the rollout as a matter of fairness: those who "game the system" should be identified and removed, while those who truly need protection should receive it.
But the deployment of this technology sits uneasily with a documented reality. An independent government immigration inspector's report from last year found that age assessments—even when conducted by trained officers using established methods like document examination, X-rays, and MRI scans—sometimes fail. The report acknowledged that in the absence of a "foolproof" test, some assessments will inevitably be wrong, and that when a child is misclassified as an adult, the consequences are severe: they lose access to the rights and protections the law guarantees them.
Human Rights Watch has called on the government to abandon the scheme entirely. The organization argues that facial age estimation technology has never been tested in refugee processing settings, only in retail and hospitality contexts like shops and bars. Anna Bacciarelli, a senior AI researcher at the group, described the approach as "deeply flawed" and "cruel and unconscionable," warning that it subjects vulnerable young people to a dehumanizing process while relying on technology whose effectiveness in this context remains unknown.
The Home Office has conducted preliminary testing on images representing different ethnicities and genders, including populations typical of asylum seekers. Those results have not yet been used to make actual decisions about anyone's status. The first live trial is scheduled to begin at Western Jet Foil, a processing center in Dover, in mid-2027. The technology will function as a supplementary tool, assisting immigration enforcement officers when a person's age is genuinely uncertain.
The decision to pursue this path follows years of sustained pressure on the UK border system. In the year ending June 2025, 111,084 people claimed asylum in the country—14 percent more than the previous year. Much of the public and political focus has centered on small boat crossings in the English Channel, creating an environment in which the government has positioned technological solutions as both necessary and inevitable.
What remains unresolved is whether a system designed to catch those attempting deception will, in practice, also catch children and wrongly classify them as adults. The independent inspector's report suggested this outcome is not merely possible but mathematically inevitable when no perfect test exists. The question facing the government—and the young people who will be assessed by this system—is whether the promise of catching more fraudsters justifies the risk of harming more children.
Citações Notáveis
Adult migrants making false age claims have exploited the system and diverted vital support away from children at risk. That is why we are rolling out AI technology to put a stop to this.— Alex Norris, Minister for Border Security and Asylum
Experimenting with unproven technology to determine whether or not a child should be granted protections they desperately need and are legally entitled to is cruel and unconscionable.— Anna Bacciarelli, Human Rights Watch
A Conversa do Hearth Outra perspectiva sobre a história
Why does the government believe this technology is necessary when trained officers are already doing age assessments?
The numbers suggest a credibility problem. In one year, 43 percent of people claiming to be children were found to be adults. The government sees that as evidence the current system is being exploited at scale, and they believe AI can catch what human judgment misses.
But the independent inspector's report says some assessments will always be wrong. Doesn't that apply to AI too?
Exactly. That's the core tension. The government is betting that AI will be more accurate than humans, but there's no evidence it will work in this specific context. It's been tested in bars and shops, not refugee centers.
What happens to a child if the AI gets it wrong?
They're classified as an adult and moved into the asylum system instead of the care system. They lose access to protections the law says they're entitled to. They're treated as someone trying to game the system when they're actually a vulnerable young person.
Is there a middle ground here—using the technology but with safeguards?
Human Rights Watch doesn't think so. They're arguing the technology itself is the problem, not just how it's used. If it hasn't been proven to work in refugee settings, deploying it on children—who have the most to lose if it fails—is the wrong place to experiment.
What's the government's response to those concerns?
They've framed it as a cost-effective solution and say the technology showed "promising performance and accuracy" in their testing. But they haven't addressed the fundamental question: what happens when it's wrong, and a child pays the price?