A predator was stalking London. He would board a crowded bus at rush hour, carrying a Metro newspaper, and sit next to a young woman. Opening the newspaper to form a curtain, he would reach over and grope her. The man first struck one summer afternoon in 2014, on the No. 253 bus in North London, grabbing the crotch of a fifteen-year-old girl. She fled the bus and called the police, but by that time he had disappeared. A few months later, in October, he assaulted a twenty-one-year-old woman on the upper level of a double-decker as it approached the White Hart Lane stadium. She escaped to the lower level, but he followed her, and he continued to pursue her even after she got off the bus. She flagged down a passerby, and the man fled. In March, 2015, he groped a sixteen-year-old on the No. 168. On each occasion, the man slipped away from the crime scene by blending into a crowd of commuters. But, each time, he left a trace, because public buses in London are monitored by closed-circuit-television systems.
When transit police played back the footage of each sexual assault, they saw the same middle-aged man in spectacles and a black parka. He had thinning hair and a dark mustache that was going gray. After consulting the electronic readers on each bus, investigators isolated one fare card that had been used on all three. If the pass had been bought with a credit card, it could be linked to the perpetrator. But the man had paid for it in cash.
The transit police found themselves in a familiar predicament: a case in which a crime is captured on video but no one can identify the perpetrator. London has more than eight million residents; unless somebody recognizes a suspect, CCTV footage is effectively useless. Investigators circulated photographs of the man with the mustache, but nobody came forward with information. So they turned to a tiny unit that had recently been established by London’s Metropolitan Police Service. In Room 901 of New Scotland Yard, the police had assembled half a dozen officers who shared an unusual talent: they all had a preternatural ability to recognize human faces.
Most police precincts have an officer or two with a knack for recalling faces, but the Met (as the Metropolitan Police Service is known) is the first department in the world to create a specialized unit. The team is called the super-recognizers, and each member has taken a battery of tests, administered by scientists, to establish this uncanny credential. Glancing at a pixelated face in a low-resolution screen grab, super-recognizers can identify a crook with whom they had a chance encounter years earlier, or whom they recognize from a mug shot. In 2011, after riots broke out in London, one super-recognizer, Gary Collins, a cop focussing on gangs, studied the grainy image of a young man who had hurled petrol bombs and set fire to cars. The rioter wore a woollen hat and a red bandanna, leaving only a sliver of his face uncovered, like a ninja. But the man had been arrested years earlier, and Collins had noticed him at the police station—in particular, his eyes. The rioter was convicted of arson and robbery, and is now serving six years.
When the transit police brought the groper case to the super-recognizers, Eliot Porritt, a detective sergeant in the unit, took up the investigation. Porritt, who is thirty-six, is rumpled and cerebral, with a mop of curly black hair. As a boy, he loved watching movies with his father, and found that he could identify actors who had been in other films they’d seen, even in tiny parts. As a police officer—first as a beat cop in Islington, and then working plainclothes on a robbery squad—he discovered that while walking the streets he could spot faces and know, in a flash, who they were, where he had met them, and whether they were criminal suspects.
Three times a week, the Met issues an online bulletin, “Caught on Camera,” featuring video stills of unidentified suspects committing crimes. Many officers ignore it, but Porritt found the activity of picking out faces quietly absorbing, like doing a crossword puzzle. He soon became known for his prowess at making identifications—“idents,” in the Scotland Yard vernacular—and last year he was asked to join the super-recognizers.
In the groping case, the transit card did not reveal the suspect’s identity, but Porritt could use it to reconstruct his movements. On a digital map of North London, he pinned a flag on each bus stop where the card had been used. He noticed that the man took as many as twenty bus trips a day. Sometimes he rode in one direction, got out, crossed the street, and went in the opposite direction. “He wasn’t commuting,” Porritt told me. “He was casing victims.”
Studying the map, Porritt plotted the various routes and developed a hunch that the man lived in Camden. Porritt grew up there, and he decided to go and ask around. He invited Alison Young, an officer who had just joined the unit, to tag along. Young is twenty-nine, with long red hair and an ebullient sense of humor. She had worked as a community-support officer for several years, but one day she was summoned to an auditorium at Scotland Yard, where dozens of officers were instructed to take a facial-recognition exam. Using a laptop, Young found matches in a series of faces that were presented like masks—without hair or other context. When the test was done, she was startled to learn that she had received the second-highest score.
By some estimates, as many as a million CCTV cameras are installed in London, making it the most surveilled metropolis on the planet. Boris Johnson, who before becoming Britain’s Foreign Secretary served as the city’s mayor, once said, “When you walk down the streets of London, you are a movie star. You are being filmed by more cameras than you can possibly imagine.”
Porritt thought that the cameras outside the Camden Road railway station might have caught the groper walking by, so he and Young visited the CCTV office there. As Porritt examined the equipment, Young gazed out a window at scores of rush-hour commuters streaming in and out of the station. Then, suddenly, she shouted, “Oh, my God. That’s him!”
Young was staring at a man just inside the entrance: he had a mustache and wore glasses. She watched him pick up a Metro from a stack on the floor and walk out of the station.
“We ran like maniacs,” Young recalled. They caught him, and after he was in handcuffs he muttered to Porritt, “I’m sorry, I’m sorry.” A fifty-six-year-old clerical worker named Ilhan Karatepe, he subsequently pleaded guilty to three counts of sexual assault and received a suspended sentence. (He was also barred from riding public transportation by himself.)
“Here was a guy that nobody could identify until Alison spotted him,” Porritt said. “Every identification we make is effectively a cold case. We’re unique. There’s no precedent for what we’re doing.”
In 2008, a postdoctoral student at Harvard named Richard Russell began working with a team of perceptual psychologists on a study of prosopagnosia, or “face blindness,” a condition in which patients are unable to recognize human faces. In extreme cases, prosopagnosia can be a socially debilitating affliction: a mother tries to retrieve the wrong child from day care because she does not recognize her own baby; a patient is shown a photograph of a woman and wonders who it is, only to be informed that she is looking at a picture of herself. But many people suffer from milder forms of face blindness, and may not realize that they are in any way abnormal. “We’re not good at talking about how we recognize faces,” Russell said. “So we assume that other people are like us.”
Until recently, only a few hundred prosopagnosics had been studied, and from this research neuroscientists and perceptual psychologists had established a binary “pathological” model: either you were normal, and could recognize faces, or you had face blindness. But new studies have indicated that although prosopagnosia can result from a stroke or traumatic brain injury, it is a heritable condition that is sometimes present from birth. It’s also much more widespread than was previously believed. With the advent of the Internet, formerly isolated individuals have found a community of fellow-sufferers.
Collaborating with two psychologists, Ken Nakayama and Brad Duchaine, Russell disseminated a bulletin in the Boston area seeking research subjects who thought that they might be face blind. The researchers heard back from many people who believed that they were prosopagnosic. But they also heard from a small group who said that they were “the opposite.”
Russell had come to suspect that facial recognition might not be simply a faculty that was either present or absent. What if it was on a spectrum? If most people are pretty good at recognizing faces and prosopagnosics are terrible at it, Russell recalls thinking, shouldn’t there be “some people on the high end”?
The team studied four people who considered themselves exceptionally good at remembering faces. First, the researchers administered a version of the exam that Alison Young aced in the auditorium at Scotland Yard, the Cambridge Face Memory Test, in which the subject is instructed to find matches among faces stripped of hair or other visual clues. A second test, called Before They Were Famous, consisted of a series of images of celebrities, from Malcolm X to Scarlett Johansson, in their youth. (The researchers chose celebrities because they needed faces that most subjects would have some prior familiarity with.)
The four subjects in the study performed extremely well. Prosopagnosics often have strange stories about how they cope with their condition. The subjects had their own curious tales about being on the other end of the spectrum. They not only recognized character actors in movies—they recognized the extras, too. In social situations, prosopagnosics often smiled blandly and behaved as if they had previously encountered everyone they met, rather than risk offending acquaintances. Russell’s subjects described the opposite adaptation: they often pretended that they were meeting for the first time people whom they knew they’d met before. After all, if you’re introduced to someone at a party and you remind him, in pointillist detail, about the circumstances of a brief meeting years earlier, he might reasonably conclude that you are a stalker. One of the subjects described an ex-boyfriend’s referring to her as a “freak of nature.”
In 2009, Russell and his colleagues published a paper in the Psychonomic Bulletin & Review, “Super-Recognizers: People with Extraordinary Facial Recognition Ability,” which concluded that there is “a broad distribution” of the facial-recognition capability. On tests, the four super-recognizers performed at least two standard deviations above the mean, which meant that they were “about as good at face recognition and perception as developmental prosopagnosics are bad.”
Earlier this summer, I visited New Scotland Yard, a tower of mirrored glass not far from St. James’s Park, to spend a week with the super-recognizers. Detective Chief Inspector Mick Neville, who created the unit, greeted me. “I’m the Henry Ford of CCTV,” he told me. “The rest are just playing at it, really.”
Neville is burly and ruddy-faced, with beady eyes and a flinty, tough-guy grin. He speaks in the gruff murmur of his native Lancashire and swathes his considerable bulk in pin-striped suits and regimental ties. As a trainee detective in Brixton, during the nineteen-nineties, Neville worked gangs, and he excelled at recruiting informants. Raised by a working-class single mother, he found common ground with the young people he arrested for robbing post offices and betting parlors. He’d slip them cigarettes or buy them a sandwich, then cajole them into snitching on their associates. In those days, CCTV images were stored in big photobooks, like family albums, and Neville would barge into a holding cell, plop the book down, and bark, “All right—names and faces!”
The first permanent CCTV cameras in London were installed in 1968, in Grosvenor Square—the home of the U.S. Embassy—to monitor students protesting the Vietnam War. More cameras were added during the seventies and eighties. In 1993, a toddler named James Bulger vanished from a shopping center in northwest England. Footage from a security camera showed two older boys leading him away. Both boys were convicted of Bulger’s murder, and the eerie images of his final walk were played incessantly on the news, solidifying the rationale for more cameras. By the late nineties, more than half of the Home Office’s crime-prevention budget was devoted to CCTV. The profusion of cameras was premised on a coercive principle: the new architecture of surveillance would dissuade people from committing crimes. “There is a friendly eye in the sky,” a Home Office minister proclaimed in 1994. “There is nothing sinister about it, and the innocent have nothing to fear.”
CCTV was clearly effective at recording malfeasance, but it struck Neville that if so many people were committing crimes on camera then perhaps the cameras weren’t much of a deterrent. He wondered why people took the risk. “Some people are drunk, or high, or just evil,” he told himself. But many rational criminals had clearly concluded that the police were unlikely to do much with CCTV footage. If you left fingerprints at a crime scene, they were entered into a centralized database. But if you left your image behind, Neville noticed, “it would just sit there.”
In 2006, he set up a dedicated unit to comb through CCTV footage and make identifications. At first, he was given mostly cops who had conditions that rendered them unfit for other duties. Sitting at a computer and scrolling through hours of surveillance footage can be migraine-inducing work; it is not for everyone. Yet Neville was stuck with an arbitrary crew of misfits. He referred to his staff, not without affection, as the Dirty Dozen.
For an official in a starchy bureaucracy, Neville has a surprising compulsion for candor. In 2008, he attended a conference and complained, “Billions of pounds has been spent on kit, but no thought has gone into how police are going to use the images. The cameras are not working.” The Guardian published an article about Neville’s remarks, highlighting his assessment that Britain’s use of CCTV was “an utter fiasco.”