Every year there’s at least one new drama on television that gets everyone talking – the “must-see”, water cooler, trending on Twitter phenomenon… 2019 was the year of BBC’s The Capture - a riveting six-part series led by Holliday Grainger and Callum Turner.
Having been acquitted of a war crime in Afghanistan, former Special Forces Lance Corporal Shaun Emery (played by Turner) finds himself accused of the kidnap and murder of his barrister. CCTV footage clearly shows the presumed victim’s last-known interaction prior to her disappearance – an assault committed by Emery at a bus stop. But is the footage all it seems?
28.03.20
It’s a fascinating premise which I expect few people will have spent time deliberating over prior to the show. The notion that CCTV footage could be unreliable - and, indeed, manipulated by government agencies - is disturbing to say the least.
There were, as you’d expect from a primetime thriller, twists and turns aplenty. But how realistic are the scenarios depicted in the show? Are we all being watched? If so, how?
We’ve all seen traditional CCTV cameras – perhaps even used the monitors on entering Homebase or Tesco to straighten out our windswept hair! These types of surveillance are in full view, but there are more discreet, undetectable and sophisticated methods out there – and opinions are divided on whether they cross a line.
The powers that be in San Francisco certainly think they do. Government agencies – including the city police and county sheriff’s department – are banned from using facial recognition technology (although federal agencies – including those at ports and airports – are unaffected). Instead, agencies are required to gain approval before purchasing surveillance tech and must publicly disclose its intended use.
Critics of facial recognition technology raise concerns that recent AI advances have transformed it into a dangerous real-time surveillance tool. On this point, I’m in two minds. Do I think the government and police should be able to track anybody at any time? No. But would I sleep easier knowing that potential threats to our society – terrorists, murderers, rapists – are being closely monitored? Absolutely.
Herein lies the quandary. Who “deserves” to be watched? And who makes that decision?
The potential for abusing such technology - as depicted in The Capture - presents an ethical conundrum of great magnitude. Real-time surveillance is only a step or two removed from microchipping, which would surely be deemed a severe human rights violation.
There is enough mistrust of law enforcement already, especially in America, for such power to be willingly granted by the public. Whilst it has been used by police forces to identify suspects and spot fraud in the past, it seems the more advanced the technology becomes, the more concern over possible misuse intensifies.
UK civil liberties group, Liberty, recently brought a case against South Wales Police to similar effect. Thought to be the first legal challenge of its kind in the world, Liberty acted on behalf of civilian Ed Bridges who took umbrage at the use of automated facial recognition (AFR) technology. The BBC reported:
Mr. Bridges brought a legal challenge after he saw a police van marked with AFR cameras as he took a lunch break in Cardiff city centre and then again at a peaceful protest at an arms fair in the city.
He argued the AFR system analysed his biometric data - digital mapping of a person's facial features - without his knowledge or consent.
The High Court ruled in September an important safeguard in the way the police force used the system was that the potential match suggested by the software was reviewed by the officer operating the equipment before the person was approached by colleagues…
…Vans have been deployed on 71 days at 39 events, with 60 people arrested at locations ranging from pop concerts to protests as well as big sporting events.
The Welsh force was the first in the UK to make an arrest using the real-time technology in 2017, ahead of the Champions League final in Cardiff.
But it emerged that of the 2,470 potential matches made using AFR, 92% (2,297) were wrong.
Arguably more important than the “should they, shouldn’t they” quandary, is the issue of accuracy. If such technology is to be deployed in the future, the results need to be infallible. That or allow chaos to reign supreme.
China has taken surveillance a step further. Trialling technology developed by a company called Watrix, Chinese police are seeking to use video surveillance footage to identify individuals by the way they walk.
In a fascinating article published in September 2019, Wired detail how the concept works in practice:
In a recently granted patent, Watrix outlines a method of using a deep convolutional neural network to train an AI system capable of analysing thousands of data points about a person as they move, from the length of their stride to the angle of their arms, and use that to recognise individuals based on their 'gait record'. Watrix claims that its systems achieve up to 94 per cent accuracy, and that it holds the world's largest database of gait records.
The vision-based methods of gait recognition being developed by Watrix and others can be used to identify people at a distance, including in crowds or on the street, in a similar way that facial recognition can – which could make it a quick and easy substitute if regulation is brought in against facial recognition. Increasingly, many video surveillance systems are collecting multi-modal biometrics. That means they may be using facial recognition and gait recognition simultaneously, which at least in theory should both increase the accuracy and tackle issues like identifying people facing away from the cameras.
I find myself shuddering at the thought of someone analysing the way I walk, but the argument that such systems could make our streets safer still rings true.
But it doesn’t stop there. China also employ AI and facial recognition technology to regulate traffic and identify drivers who violate road rules. In Shenzhen, jaywalkers are publicly named and shamed on large LED screens, with a fine issued via instant messaging thrown in for good measure.
Interesting, then, that a number of China-led initiatives are gathering momentum globally. Some of the country’s largest tech firms – including Huawei, Hikvision and ZTE – are looking to expand their company footprint through the export of AI-powered surveillance solutions. Of the 63 countries they currently supply, 36 nations have signed on to China’s “smart city” infrastructure project – the Belt and Road Initiative.
China’s vast surveillance state reportedly extends to over 172 million cameras – roughly three times as many as are operating in the US – accounting for 43% of a global $47 billion business. The expansion of this empire has sparked serious concern overseas. In particular, the possibility of data being siphoned back to China, thereby strengthening authoritarian regimes and leading to increased influence abroad for the Chinese Communist Party.
Data security is rightly becoming a high priority for individuals, companies and governing bodies around the world. It is a scary thought not knowing who has access to your personal information – or how they may use it. Funny then, that I find myself even more disturbed by reports that surveillance cameras are in action in the public toilets next to Beijing’s Temple of Heaven to prevent toilet paper theft!
In March 2017, Fortune reported:
High-definition cameras installed at the city’s Temple of Heaven restrooms capture facial images and then reportedly prevent any one person from taking more than 60 centimeters (or about 23 inches) of paper within a nine-minute period… If the same person appears during that period, the toilet paper dispenser stops working.
Perhaps having strangers analyse the way I walk is preferable after all…
