The Home Office has stated that its implementation of face scanning technology is intended to serve as an “additional” resource for border personnel, emphasizing that it will not “supersede or undermine human judgment.” However, the department has not clarified how this technology will be applied in practical situations. According to a spokesperson, “individuals will always be regarded as children until a further evaluation is completed” in cases where uncertainty arises.
Revised Timelines for Implementation
In July 2025, the UK government initially revealed intentions to incorporate facial age estimation alongside the assessments made by border staff to evaluate migrants. The rollout has since been postponed to 2027, with the Home Office asserting that it will utilize “advanced AI technology” to combat false claims and prevent “adults from exploiting the system.”
Over the last five years, AI facial recognition technology has become integral to age verification initiatives. These measures have been enforced by lawmakers across social media platforms, adult websites, and selected retailers to ensure user age compliance. Additionally, selective trials have been conducted at various bars and stores within the UK. Facial age estimation analyzes distinct facial characteristics, utilizing models trained on millions of age-labeled images to generate an estimated age. High-performing algorithms in controlled tests can approximate a person’s age within 2.5 years.
However, the efficacy of these algorithms can be inconsistent, influenced by diverse variables such as gender, demographic background, and image quality. For instance, low-resolution images or poor lighting can significantly hinder system accuracy. Reports indicate that certain algorithms have been misled by images of fictional characters, indicating vulnerabilities in the technology. Despite these concerns, the Home Office has proceeded with its program.
An April 2025 document from the Home Office, created prior to the acquisition of face-scanning technology, examined seven facial age estimation algorithms with a pool of over 2.5 million images. The internal report reveals that the yet-to-be-named “top-performing algorithm” displayed “significant inaccuracies” when assessing images of individuals from Sub-Saharan African backgrounds. Alarmingly, this algorithm often misclassified a 17-year-old as over 18, particularly misrepresenting female subjects.
Each year, tens of thousands of individuals submit asylum applications in the UK, many arriving after perilous journeys across the English Channel in small boats. Currently, border staff evaluating the age of those claiming to be under 18 base their initial assessments on physical appearance, responses during interviews, and overall demeanor. Such determinations occur during the “first encounter,” according to Home Office guidelines. Official statistics show that since 2010, approximately 40 percent of those subjected to age assessments have been classified as adults.
The leaked report suggests that the findings stem largely from tests using high-quality images of documented individuals, which may imply even lower accuracy in real-world applications. The Home Office has indicated that facial age estimation technology is intended to aid immigration officials at the point of initial assessment, although the reliability of these technologies remains under scrutiny.