The UK government is turning to algorithms to solve a persistent border dilemma. Starting next year, officials will begin trialing AI-driven facial recognition technology to estimate the ages of asylum seekers, a move designed to identify adults who may be misrepresenting themselves as children.

This is a high-stakes pivot. The Home Office has awarded a £322,000 contract to Harlow-based Akhter Computers Ltd to develop and test the software. By mid-2027, the system is expected to be fully operational at border processing centers, including the Western Jet Foil facility in Dover. The goal is simple: stop adults from gaming the system. The reality is far more complex.

The Cost of Misclassification

For the government, the stakes are measured in resources and security. In the year ending March 2026, over 6,400 migrants claimed to be children at the border. Subsequent assessments found that 43 percent were actually adults. These individuals, if misclassified, enter the care system, placing significant strain on local councils and diverting support intended for vulnerable minors.

Minister for Border Security and Asylum Alex Norris argues the technology is a necessary intervention. "Adult migrants making false age claims have exploited the system," Norris said in a statement. The Home Office maintains that the software will provide an objective, cost-effective layer of scrutiny to support border officials who currently rely on subjective observations of appearance and demeanor.

A Technology Under Fire

Critics are not convinced. Human Rights Watch has labeled the initiative "cruel and unconscionable," warning that the technology remains unproven. The core fear is that an algorithmic error could strip a genuine child of their legal protections, forcing them into adult asylum accommodation or detention.

The British Association of Social Workers (BASW) shares these concerns. They argue that age assessment is a nuanced, human-centric process that cannot be reduced to a photograph analysis. "This important task should not be open to shortcuts," said Professor Sam Baron, interim CEO of BASW. The risk of a false positive is not just a statistical error; it is a potential safeguarding catastrophe.

The Limits of the Data

While the Home Office claims "promising performance" from initial testing, the technology has yet to face the pressures of live, high-stakes border operations. Previous reports from the independent immigration inspector have highlighted that even human assessments are far from foolproof. The inspector noted that both adults have been wrongly classified as children and children as adults, creating a cycle of systemic failure.

Can an algorithm bridge this gap? The Home Office says yes. They have tested the software against diverse ethnic and gender groups to ensure accuracy. However, the lack of transparency regarding the software’s error rates remains a point of contention for civil liberties groups.

Key Takeaways

  • The Home Office will trial AI facial estimation at the Dover processing center next year, with a full rollout planned for mid-2027.
  • Official data shows that 43 percent of migrants who claimed to be children in the year ending March 2026 were later assessed as adults.
  • Human rights groups and social workers warn that the technology is unproven and risks denying vulnerable children their legal protections.

The first live trials at the Western Jet Foil facility will serve as the true test. By the time the system reaches full deployment in 2027, the Home Office will need to provide concrete evidence that the software’s accuracy exceeds that of human social workers. If the error rates remain high, the government will face a choice: abandon the tool or defend a system that could systematically misidentify children at the border.