Met Police 'Hamstrung' As Knife Crime Epidemic Fuels Calls For Facial Recognition

Police officers search near where a 16-year-old boy was killed, on November 07, 2018 in London. (Photo by Dan Kitwood/Getty Images)Getty

Bleak headlines in London last week, where five people were killed by knives in five separate incidents over the course of six days. Three of the five were teens, growing up in a city where for increasing numbers of kids, carrying a knife has become as normal as carrying a cell phone.

Stabbings will always cede the global headlines to the relative efficiency of gun crime. And, the same week, the latest mass shooting in the US claimed many more lives. But it is the relentless normality of knife crime amongst elements of London’s young that is the story here. Teens killing teens outside fast food joints in a city where the advice to under-16s is that the most dangerous time of day is the journey home from school.

For London’s Metropolitan Police Service, the new normal is reassuring a city frightened by escalating violence, whilst engaging with politicians distracted by Brexit and factional infighting, and doing so with tighter budgets, translating to fewer officers out on the street. The inevitable answers are active intervention and new technology, but that prompts sensationalized headlines and even a legal challenge from a privacy lobby that wants the public to be kept safe (one assumes) but remain unwatched. And so the capital's knife crime epidemic has become the embodiment of its thin blue line, stretched ever thinner by increasing workloads, decreasing resources and politics.

The Reality Of The Situation

Cressida Dick, the Metropolitan Police Commissioner, has been impressive and forthright in her response, promising to pursue the 190 plus gangs in London behind the worst of the violence. She has not shied away from discussing, frankly and unsettlingly, the realpolitik of the situation. Unclear direction and lack of leadership from lawmakers. Political hand-wringing. Budget decisions taken in recent years, which now hang in the air, awkwardly and provocatively. Unsurprisingly, Home Secretary Sajid Javid’s call for police to “step up” did not land well, set against this backdrop of police cuts and priority stretch.

Home Secretary Sajid Javid and Metropolitan Police Commissioner Cressida Dick during a visit to Brixton, South London. (Photo by Gareth Fuller/PA Images via Getty Images)Getty

If you want to tackle knife crime, you really need to stop and search those you suspect might be carrying knives, and thousands of potential weapons have been pulled from the streets in this way. But ‘stop and search’ fosters controversy: go to any major city in the world, and targeted policing will always risk the accusation of some form of bias or discrimination.

Beyond counter-terrorism, which has its own legislation, UK police can stop and search a person they suspect to be carrying drugs or weapons or both, or they can stop and search a person in a location where there have been or are considered likely to be “incidents involving serious violence.” Searches, where a person is thought to be carrying an item, have fallen in recent years, but locational searches are on the rise. The fact is that much of the knife crime epidemic is localized, territorial and gang-related.

Cue Facial Recognition

This is where facial recognition comes in. Systems that can memorize the faces of persons of interest. Networks of gang members. Wanted criminals. Those suspected of involvement in serious violent crimes. Systems that don’t need prior personal engagement to recognize an individual. Systems that don’t rely on context, that can be set up to deliver specific operational outcomes, that can be localized and needs based, that can continually improve. This is not simply software. This balances the delivery of successful outcomes with the technology platforms that underpin it. People are in the loop every step of the way, the final decisions are taken by those people, not by their machines.

The best facial recognition systems can be tailored to meet specific requirements, operating in real-time on connected cameras, as opposed to offline on recorded surveillance footage. Operating live, meaning networking instant watchlist updates and results. And operating across multiple platforms at the same time: this allows for verification of results and a significant increase in effectiveness. Imagine the potential of a network of CCTV, vehicle cameras and bodycams, all updating their watchlists and enrolments in real time as an intelligence picture changes after a serious incident. Imagine verifying results on a second platform, or even with a second and third AI engine, all in real-time. The technologies are built around machine learning platforms. What they can do today will be materially surpassed in weeks, never mind months and years.

Automated facial recognition system. (Photo credit should read MARKUS SCHREIBER/AFP/Getty Images)Getty

Earlier this year, Cressida Dick told a London Assembly committee that facial recognition “is getting better and better and better by the minute… if there is a technology that we can use lawfully… to identify against a small list of wanted offenders for serious violence, I think the public would expect us to be thinking about how we can use that technology and seeing whether it is effective and efficient for us.”

The Met is currently trialing facial recognition, assessing its effectiveness in “the prevention and detection of crime by identifying wanted criminals.” The trials have been extensively covered in the media. They have been structured to encourage public feedback and debate. As Ms. Dick told the same committee meeting, “it is a really useful function for the Home Office to make sure the Government is engaging in debates with the public about the balance between privacy and security in the light of changing technologies. That needs to be done dynamically and quite quickly.”

A Serious Need For Clarity

But it hasn't happened dynamically or quickly. And at the end of last week, in an interview with the Daily Telegraph, Ms. Dick described her officers as being “hamstrung” by the lack of political regulation and direction for the “much-needed use of facial recognition” on the frontline. The vacuum left by this lack of regulation is critical because it is being filled by misleading claims about the way the technology works and how it is being used.

This testing, and similar trial deployments by other police forces in the UK, without the regulatory or legal frameworks the police have asked the government to produce, has fueled a backlash from the privacy lobby. The campaign groups have described facial recognition surveillance as “an inherently authoritarian tool for population management,” as threatening “privacy, freedom of expression and right to protest,” as “checking millions of people against police databases,” and as “dystopian policing.” In June, two of the lobby groups, Liberty and Big Brother Watch, launched a legal challenge against the use of facial recognition by two police forces: the Met and South Wales. The challenge remains ongoing.

These same debates are taking place worldwide. Also last week, Andy Jassy, the CEO of Amazon Web Services, defended the company’s controversial facial recognition technology in an internal meeting, where there had been discomfort amongst employees around sales to law enforcement agencies. He said: “Rekognition is actively used to help stop human trafficking, to reunite missing kids with parents for educational applications, for security and multi-factor authentication.”

The leading versions of facial recognition, when properly trained and deployed, are exceptionally accurate. That is not up for debate. The real debate should be the usage of the technology. Who is on a watchlist, and why? What regulatory framework is in place? The lobby groups' scaremongering prevents just such a constructive debate taking place. And that is dangerous. Facial recognition should not be everywhere, nor should it look for everyone. It should be deployed only where it is appropriate and proportionate to do so. One can assume that most Londoners value the lives of the city’s kids highly enough that if a technology can make a tangible difference, it should be trialed and adapted to see how it can be used most effectively.

Tackling The Knife Crime Epidemic

Behind the testing and development of facial recognition is the opportunity to deploy a technology ecosystem that can provide new capabilities to frontline policing in a manner that society deems right. Let’s remember, these same biometric technologies are being used to tackle international and domestic terrorism, people and drug trafficking, pedophile networks and their victims, children and vulnerable adults missing from home. How many missing kids in the UK pass through one of London's main rail or bus stations each month? Facial recognition with a watchlist of missing kids could be deployed for just such a use case.

Crime hotspots, known networks of gangs and affiliates, wanted criminals, persons of interest seen by CCTV in the vicinity of a violent crime, targeted operations to take knives from the street, all can be supported by a consistent use of facial recognition. A stop and search, where the police officer has access to a connected mobile device or bodycam with a facial recognition capability, can only help support the justification for such operations and their effectiveness. Again, secondary verification, appropriate use of watchlists, regulations around data retention, and a final decision on any engagement or intervention taken by a person, not by a machine, should all be part of the operating model.

The scene where 15-year-old Jay Hughes was stabbed and killed, on November 06, 2018, in London. (Photo by Dan Kitwood/Getty Images)Getty

It is 2018, the era of AI and intelligent video. And in London, we have a situation where lives can be saved by these technologies. Facial recognition works. It can make people safer. And so it’s time to set some regulations and start making use of it. The Met Commissioner has shown commendable leadership in setting out requirements to move past the endless hand-wringing and tortuous debate; in essence: don’t plead with me to solve the problem, whilst tying one of my hands behind my back.

https://www.forbes.com/sites/zakdoffman/2018/11/10/met-police-hamstrung-as-knife-crime-epidemic-fuels-calls-for-facial-recognition/#4665dfa974d3