344
Productivity & Workflow355
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative169
Customer Support131
Photography & Imaging156
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Leaked Home Office tests found facial age estimation can misclassify children as adults and performs worse for some groups. The UK still plans a rollout.
In short: The UK Home Office plans to use AI that estimates age from a face photo at the border, even though internal tests found frequent mistakes and possible bias.
The British government is preparing to use “facial age estimation” to help decide whether some asylum seekers are under 18. Facial age estimation is an AI tool that looks at a person’s face in a photo and guesses their age (like a digital bouncer that tries to judge age by appearance).
WIRED and Lighthouse Reports obtained an internal Home Office report on tests of seven of these tools. The report says the best-performing system still made serious errors, including regularly mistaking children for adults. It also found performance was worse for some groups, especially Sub-Saharan Africans, and worse for females. For female Sub-Saharan Africans, the average error was 4.6 years, which could mean a 13.5-year-old being guessed as 18.
The Home Office first announced the plan in July 2025 and has delayed rollout until 2027. The department says the tool will be “additional” and will not replace human judgment. It also says that when there is uncertainty, people will be treated as children until a further assessment is done.
The report warns that real-world conditions could make results worse. Photos taken at first encounters are often low quality, and stress and trauma can affect someone’s appearance, which may hurt accuracy.
Whether someone is treated as a child or an adult can change where they are housed and what legal protections they get. If a system wrongly labels a child as an adult, the consequences can be severe, including placement in adult-only detention settings. That is why accuracy and fairness, including how the tool performs across different backgrounds, matters so much in this use case.
Source: Wired