Facial Recognition Software Error Leads to Wrongful Arrest in UK

Alanbatnews -

British authorities recently arrested a man at his home due to a flaw in facial recognition software, which incorrectly placed him at the scene of a robbery he never visited, according to reports.

Alfie Chouhdry, 26, was detained for over ten hours before the error was discovered and he was released. Chouhdry is now seeking compensation.

Police relied on facial recognition technology developed by the German company Cognitec. The software identified Chouhdry as a suspect in a robbery that occurred more than 160 kilometers from his residence. Discrepancies existed between Chouhdry's appearance and that of the suspect seen in available footage.

The UK police force relies heavily on facial recognition software to identify suspects, conducting over 25,000 searches monthly within a database of suspect images containing as many as 19 million photographs.

Facial recognition software struggles to accurately detect and record facial features among various minority groups. Error rates can reach as high as 5%, compared to approximately 0.04% for individuals with lighter skin tones, according to reports.

This is not the first time Cognitec's software has led to a mistaken arrest. Police in New South Wales, Australia, completely suspended the software's operation in April 2025, according to reports.

The suspension was attributed to the software's outdated algorithm, which had not been updated in over 11 years, making its results unreliable.

In a prior incident, the technology led to the wrongful arrest of a 14-year-old boy after incorrectly identifying him at a crime scene.

Cognitec states on its official website that it updated its facial recognition software and algorithms in February.