Vans fitted with life facial recognition technology being used by police forces across Britain. Photo / UK Govt Home Office
Half a second after the live facial recognition (LFR) camera snaps the young man walking towards Leicester Square, an alert pings on Sergeant Kevin Brown’s handheld screen with a matched image.
Two faces flash up alongside each other on his smartphone-sized device.
One is an image of the man taken by the LFR camera and the second is a police mugshot with details of the serious offences he is being sought for by Scotland Yard.
The ping and images have simultaneously registered on 10 other police officers’ screens, alerting them to the match from one of the five LFR cameras deployed in a pedestrianised street leading to the square in central London.
Their task is to identify the man caught by the camera on the busy street.
Once found, the cameras can then confirm whether he is one of 16,000 suspects on a police watchlist generated the night before.
Within seconds, the young man in his 20s is stopped and questioned by two officers. In minutes, they establish he is wanted by the Met for a crime meriting at least a year in jail. He is arrested, handcuffed, and taken into custody.
For legal reasons, Brown, a surveillance specialist now leading the LFR deployment in London, cannot disclose his offence, but says: “It’s someone who should not be walking the streets and, as a result of this, they are no longer walking the streets”.
Brown is an advocate for the technology, seen as part of a Big Brother state by some, which he says is a modern “intelligence” tool that’s making policing more effective and efficient.
This summer, the LFR cameras deployed at Notting Hill Carnival caught Tabsart Abderahmen, 58, who had been on the run for 10 years after sexually harassing a woman.
Wanted by the courts since October 2015, he was finally arrested by the Met after his image was confirmed to be on the watchlist.
The technology – pioneered by the Met and South Wales Police – is about to be rolled out nationwide backed by the Home Office, starting with seven constabularies who will get 10 new vans fitted with LFR cameras.
The operation last Thursday night local time coincided with the Met Police’s publication of its annual LFR report, showing that twice-weekly deployments of cameras in just four locations have garnered 2077 alerts leading to 962 arrests over the past year.
Nearly two thirds of those arrests – some 549 – were people wanted by the courts, 347 were sought by the Met for crimes including rape, theft, and assault, and 85 were for a breach of court orders, including registered sex offenders and stalkers.
More than a quarter of the 962 arrests were for individuals involved in violence against women and girls, including those suspected of strangulation and domestic abuse.
In the year to September, LFR cameras scanned 3.1 million faces in seconds, measuring metrics like the distance between people’s eyes, nose and mouth, creating in split seconds a “digital signature” of the face. Each was processed and cross-referenced by an algorithm against a watchlist.
Those in baseball caps, beanies and balaclavas can still be identified. When the computer throws up a match, officers take over and approach the individual, treating the images as “intelligence” to be checked. They can override the “match” if they regard it as inaccurate.
The annual report shows that out of the 3.1 million images, there were 10 false alerts where the “match” proved to be the wrong person. Eight were black or ethnic minorities, potentially reflecting lingering evidence of racial bias in the technology.
However, in four of those 10 cases, they were discarded by officers before the individual was engaged. In the six other cases, the person was approached by police and spoken to for under five minutes before no further action was taken. The Met points out this is an error rate of just 0.0003 per cent, or one in 300,000.
Hi-tech cameras and algorithms are also not immune to human fallibility. Last week officers mistakenly released one of the suspects that had been “matched” to the watchlist, to the frustration of Brown.
In the 1.5 hours observed by the Telegraph, they had two arrests, one accidental release, one subject to a court order and one “not located”.
Brown, who joined the Met at 19 and has served for 22 years, explained spotting matches in the dusk light was challenging due to the sheer volume of people in Leicester Square.
However, he had no doubts that LFR is a potential game-changer. “It is incredibly precise, really accurate and at a time when resources are stretched, anything we can do to make policing more efficient and move with the times has got to be positive,” he said.
The Met is preparing for deployments of LFR vans this year and has introduced the first static LFR cameras in Croydon.
Britain, the United States and China are the most advanced users of LFR, which the Met says is supported by four in five Londoners, according to polls.
Deployment is targeted at crime hotspots and signs warning the public they are being filmed by LFR cameras are placed all around the van’s location. Any images captured of people not on the watchlist are irretrievably deleted instantaneously.
Sign up to Herald Premium Editor’s Picks, delivered straight to your inbox every Friday. Editor-in-Chief Murray Kirkness picks the week’s best features, interviews and investigations. Sign up for Herald Premium here.