Facial recognition technology used by London's Met police at the Notting Hill Carnival for the second consecutive year was unable to tell the difference between a young woman and a balding man.
The system, which was said to be "top-of-the-line", was being used to spot so-called troublemakers, according to the security company Sophos which quoted a rights group worker who had been invited to view its operation.
The carnival, held in the last week of August each year, attracts about two million visitors.
The company said the system had proved to be useless in 2016 and this year it was worse than that: "it blew up in their faces, with 35 false matches and one wrongful arrest of somebody erroneously tagged as being wanted on a warrant for a rioting offence".
{loadposition sam08}Silkie Carlos, the technology policy officer for British civil rights group Liberty, said the system had "all the hallmarks of the very basic pitfalls technologists have warned of for years – policing led by low-quality data and low-quality algorithms”.
But the Met police viewed the use of the system as a resounding success, Carlos said, because it had come up with a solitary positive match.
Sophos said even this one correct match had some issues. "Even that was skewered by sloppy record-keeping that got an individual wrongfully arrested: the automated facial recognition was accurate, but the person had already been processed by the justice system and was erroneously included on the suspect database."
The company said that of 454 people arrested at last year's carnival, the system was unable to tag a single one as a potential troublemaker.
Photo: courtesy Adobe