The UK government will start using AI to estimate the age of asylum seekers at the border. This technology scans faces and predicts age. It’s supposed to help officials decide if someone is a child or an adult.

The plan is set to launch in 2027 despite internal reports showing the technology is unreliable and biased. Tests revealed frequent errors, especially with Sub-Saharan African migrants, the largest group crossing into the UK. In some cases, young girls were misclassified as adults by nearly five years.

Misclassifying children as adults strips them of legal protections. They can be detained with adults, risking abuse and trauma. The stakes are life-altering. Yet the Home Office insists the AI tool will only assist human judgment, not replace it.

Experts and rights groups have condemned the move. Over 60 organizations—ranging from Amnesty International to Human Rights Watch—warn the AI system is discriminatory and inaccurate. They highlight that the system struggles most around the critical 16-to-18-year-old boundary, where legal status shifts.

These flaws stem from biased training data and poor photo quality at border crossings. Stress and trauma from dangerous journeys may also age a child’s appearance, skewing AI predictions. The government’s own tests confirm these issues but still pressed forward.

The Home Office disbanded a scientific advisory committee on age assessments just before announcing the AI rollout. Former members say they were never consulted and suspect the committee was shut down to avoid criticism.

The AI supplier, Cognitec Systems, was contracted after testing seven algorithms. Their system performed worst on photos of African migrants, misclassifying more than half of 16-year-olds from West Africa as adults. This demographic forms the majority of recent arrivals.

Despite the controversy, officials claim the technology modernizes age checks and helps prevent fraudulent claims. They say individuals will be treated as children in cases of doubt until further assessments occur. Still, the exact real-world use and impact remain unclear.

Using AI to judge age at border crossings sets a dangerous precedent. Other countries watching may adopt similar systems, potentially harming vulnerable migrants worldwide. When human lives hinge on flawed algorithms, skepticism isn’t just prudent—it’s necessary.