Skripal Case Shows the Limits of Surveillance

(Bloomberg Opinion) -- The U.K.’s ubiquitous surveillance cameras have clearly played an important role in the attribution of the attempted poisoning of an ex-spy in Salisbury in March to the Russian military intelligence. Thanks to the cameras, the two Russian suspects’ movements were tracked exhaustively. But this seeming success also lays bare the biggest problem with universal surveillance: If everyone is tracked, no one is, so the cameras can only perform their function so late after the fact that even those criminals who are identified are less likely to be apprehended.

It’s estimated that London is watched by 500,000 CCTV cameras, one for every 16 residents. (Many U.S. cities are brimming with cameras, too, despite the protests of privacy advocates.) That made it possible for U.K. police and intelligence to recognize the men identified as Alexander Petrov and Ruslan Boshirov (almost certainly not their real names) as they passed through the airport, the London underground, a dingy hotel and the town where they allegedly smeared former colonel Sergei Skripal’s door handle with a military-grade poison which failed to kill the ex-spy and his daughter – but took the life of a women who accidentally came into contact with the substance.

The exercise has taken months, though. It involved scanning through a mountain of footage and using not just sophisticated recognition software but human “super-recognizers,” people with an exceptional memory for faces. It couldn’t be otherwise: The cameras, obviously, don’t conform to any unified standard. According to this year’s Video Surveillance Report from Ifsec Global, a British website that covers the security industry, only 16 percent of the installed video security systems use artificial intelligence-based analysis of the footage; the survey done for the report revealed that 74 percent of security systems use at least some analog cameras. A quarter of the systems were installed six years ago or more.

Academic studies of the efficiency of camera surveillance show that they have the potential to reduce property crime, like pickpocketing and other petty theft, though not violent crime. Cameras also tend to work best if they’re actively monitored and police react quickly to the crimes they capture. But whatever their actual effect, they have failed to prevent an explosive growth of serious crime (shootings, knifing, rapes) registered in London in recent months. As for petty crime, cameras are supposed to have a psychological deterrent effect, but small-time criminals have read the same newspaper stories as everyone else, citing the London Metropolitan Police’s recently updated crime assessment policy: If the damage is less than 50 pounds ($65), the police won’t bother investigating, nor will they pursue shoplifters or the perpetrators of “low-level assault” if they’re required to watch CCTV footage for more than 20 minutes.

Berlin’s number of security cameras is smaller than London’s by an order of magnitude. In 2016, the city government knew of 14,765 of them, one per 235 city residents; the number has likely increased: Immediately after the terror attack on a Christmas market at the end of that year, its perpetrator managed to slip away from the police and there was no footage by which he could be tracked, so the federal government called for more surveillance. But Germans are still leery of it given the country’s not-so-distant history of two distinct surveillance states, and resistance is still powerful; Berlin’s governing leftist parties promised in their coalition agreement last year not to expand surveillance.

Despite this powerful pushback, Berlin is a safer city than London. For example, 32,869 robberies took place in London in the 2017/18 financial year (which ended in March), about 402 per 100,000 population, while in Berlin, there were 4,242 robberies in calendar 2017, some 115 per 100,000 residents.

Obviously, that doesn’t prove anything. The difference in crime rates has to do with much more than just a city’s preferred surveillance techniques. But the persistence of this difference and the opposite directions of the trend – crime is growing in London and falling in Berlin – show that the ubiquitous cameras are likely not the answer to crime concerns.

Being watched all the time may be comforting in a way. Police will only examine the footage when there’s a serious case, like a foreign chemical attack. Other people are recorded and forgotten, as anonymous as if no camera eye ever peeked at them. That, however, depends on the watchers’ discretion. China is building a comprehensive video surveillance system meant to cover all “key public spaces” and industrial locations by 2020; it’s already got 170 million CCTV cameras, most of them more modern than in the U.K., and it’s getting 400 million more. That’s not an effort to deter pickpockets but rather to be able to watch and analyze data on any citizen.

The universal surveillance in the U.K. did allow police to establish the suspects’ identities, but important as that is for the U.K. government’s Russia policy, the identification came too late to actually catch them. Nothing will make them travel to the U.K. again.

Is this outcome worth recording millions of innocent citizens and subjecting them to the roughly 1 percent risk of a false positive that exists even with state-of-the-art face recognition technology? Is it worth the risk of surveillance system hacks, which are way too easy to carry out? I doubt it.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Leonid Bershidsky is a Bloomberg Opinion columnist covering European politics and business. He was the founding editor of the Russian business daily Vedomosti and founded the opinion website Slon.ru.

©2018 Bloomberg L.P.