‘Shaking from head to toe’ — How the facial recognition blame game leaves shoppers stranded
Inside the ‘Kafkaesque nightmare’ of clearing your name
Mary's husband felt "isolated" after being wrongly accused of shoplifting at Home Bargains. The experience left him feeling "scared to go into a shop". Another woman was so frightened by a similar incident at Budgens that her husband was "shaking from head to toe". Both were innocent victims of the facial recognition system.
Facial recognition technology scans a person's features and processes biometric data to determine whether they should be allowed entry to a space. While it may occur instantly, the consequences can be long-lasting and deeply personal.
In a meeting at the House of Commons, people who were falsely accused by facial recognition technology discussed their experiences with Labour MP Dr Allison Gardner. Big Brother Watch, a civil liberties group, sponsored the event to give people a chance to voice their grievances. However, neither Sainsbury's nor Home Bargains sent representatives despite being involved in the technology's implementation.
The testimonies shared by the wrongly accused people had many similarities. A normal activity was turned into a public, confusing ordeal where they had to defend themselves against allegations they did not understand. Attendees felt "bewildered", "isolated" and "confused".
One woman was wrongly intercepted simply because the system had identified the people walking in directly behind her. Algorithms, bias, poorly configured training data, and redress mechanisms were discussed. The attendees felt a bureaucratic exhaustion, having no clear path to clear their names.
Big Brother Watch's head of advocacy, Jack Coulson, said, "You can taste the indifference." When asked about liability for errors, Facewatch, the technology provider, stated that both the retailer and the company share responsibility, but legal liability depends on the facts and cannot be determined without a thorough investigation.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.