Facial Recognition Unmasked
Face coverings are a fact of life in the pandemic. It is mandatory to wear them in shops, schools and on university campuses. Slowly they’re becoming part of prison life too.
Recently masks have sparked a fresh debate about police surveillance.
Elliot Tyler, a Longford scholar who is a cyber security and forensics student at Portsmouth University has been delving into the controversial policing technology…
For a long time, I have been wary of facial recognition technology – software capable of matching a human face, either from from a digital image or a video frame, against a database of ‘wanted’ faces – used without people’s consent. I moved away from my London birthplace three years ago but returned to the city every so often to meet and have a drink with school friends. It always concerned me that my evening could be brought to a halt at any point by police officers scanning my face and determining I was somebody ‘wanted’, which I most certainly was not, having paid any debt to society that I’d previously owed.
Despite being very familiar with the ways of the police through my work as a police cadet, I did not trust them to exercise their duties impartially, responsibly and fairly. Little did I know that before long a global pandemic would see people hiding their faces, potentially disrupting the captured images.
Broken Images
In the summer, a document leaked to an American publication outlined the worries of US agencies that face masks, now commonplace due to the COVID-19 pandemic, could ‘break’ facial recognition. The document, released as part of what became known as the ‘BlueLeaks’ hack, emphasised the possibility of ‘violent adversaries’ using protective masks to evade biometric identification algorithms – in other words, donning a mask to cheat identification.
Now that coronavirus has changed the world, and face coverings routinely cover the mouth and nose – key distinguishing features for anyone – the debate has taken on a new dimension. Commentators from both sides of the debate have raised concerns about how the wearing of masks could affect the accuracy of the technology.
To this day, it remains unclear, due to the regular emergence of new, conflicting survey results, whether the general public supports this new policing initiative. Opponents of the technology suggest there is a risk to citizens’ privacy. In a free world, individuals are supposedly allowed a choice when it comes to matters of consent. However sceptics observe that permission can be simply non-existent when it comes to facial recognition. And anyway, importantly, those accused of crimes still have rights. Supporters claim if you go about your daily life in a law-abiding way, you have nothing to fear and everything to gain from effective identification to keep our streets safer.
What the experts say
I took it upon myself to consult three facial recognition experts with different perspectives on this hotly debated policing technique. First up, Chief Superintendent Paul Griffiths, the President of the Superintendents’ Association.
‘The public needs to be kept safe,’ Griffiths said to me firmly. ‘And that is achieved using CCTV, ANPR (number plate recognition), speed cameras, and other surveillance technologies.’
Our conversation continued with me citing US academics, who claim that effective facial recognition technology can prevent false arrests by quickly and accurately identifying faces.
‘It certainly isn’t the only method we rely on.’ He was keen not to be too gung-ho. ‘Data involves responsibility,’ I was told. ‘We need to be satisfied that the use of any data can support the police in their goals.’
The list of benefits from effective use of facial recognition, according to Griffiths, are early detection of ‘wanted’ individuals, allowing the efficient scrambling of police resources so officers can secure themselves and the public, possibly saving lives. Police officers can, therefore, spend their time maintaining order on the streets instead of searching aimlessly for suspects. It was explained to me that developments in technology should be embraced by police forces, but only where its use is necessary and proportionate. Police will operate with scrutiny, accountability and oversight when using personal data, Griffiths emphasised.
Lord Blair, Longford Lecture 2019
As a brief side note, in November 2019, at the annual Longford Lecture, I paid close attention to a delivery by Lord Ian Blair, former Commissioner of the Metropolitan Police from 2005 to 2008. Lord Blair explored the topic of ‘Where Next for Policing and Criminal Justice’. Concerned about a ‘tattered’ justice system, he promoted the wider roll-out of body worn cameras for future effective and consensual policing. Yet, surprisingly, he neglected to mention the ever-emerging matter of authorities’ use of facial recognition technology. The omission still intrigues me – but that’s one for another time.
Back to my second expert in this fraught field. Richard Lewis, recently retired Deputy Chief Constable of South Wales Police held a similar view to Griffiths. ‘Facial recognition can be a powerful technology for crime detection and prevention,’ he told me, adding, ‘when used appropriately.’
South Wales Police were this summer subject to legal action, brought by a father of two, Ed Bridges, who objected to his image being captured on a lunch break in Cardiff City Centre and again at a peaceful protest. With the support of the campaign group Liberty, the Court of Appeal found that the specific uses of facial recognition were unlawful.
But the Court also found its use was a ‘proportionate’ interference with human rights, as the benefits outweighed the impact on Bridges.
After the ruling, South Wales Police, who have used this type of identification method at big sporting fixtures, concerts, and other large events since 2017, said they could work with the ruling. A ‘factsheet’ produced by South Wales Police, which was sent to me before the Court decision, shows that in 2019, facial recognition technology resulted in twenty-two arrests and disposals at Welsh music and sporting events. It also rebuts common concerns about gender or racial bias within the technology. Typically, black men are thought to be disproportionately picked out.
The future of the surveillance business
Whatever the effect of face masks on the camera technology, I expect civil liberty campaigners will continue to voice their concerns about the premise of this policing technology. London’s Metropolitan Police is said to be the largest police force outside of China to use facial recognition, dubbed an ‘authoritarian mass surveillance tool’ by Big Brother Watch. Their spokesperson told me that public spaces are being turned into biometric surveillance zones, without any clear legal basis or authority, and contrary to the police rationale. They emphasised concerns about biased targeting of people of a certain ethnicity or demographic.
Be under no illusion, surveillance is big business. At the start of the year it was estimated that by 2024, the global facial recognition market would generate £5.5 billion of revenue. Of course, in the new post-COVID world, that may no longer be the case.
So where does this contentious technology head in the policing of tomorrow? In England and Wales, the police’s technology is still in a developmental stage, with three universities currently working with the Home Office to improve recognition accuracy.
This is by no means the end of the debate; in fact, I would say it’s merely the beginning.