All Commentary
Monday, August 29, 2016

Surveillance, Not Facial Recognition, Is the Real Problem

Facial recognition merely intensifies the problem of surveillance because we can no longer be lost in the crowd.

The state of New York has increased its DMV facial recognition technology, leading the governor to brag that 100 new cases of identity theft have been solved and 900 unsolved cases have been opened. This might immediately conjure up sci fi images of a wrongfully-accused protagonist being hounded by facial recognition technology as he navigates a hostile city. But how much should we really fear this technology?

With the exponential nature of technology growth, facial recognition will likely become a ubiquitous part of our lives.

New York’s DMV is only using facial recognition technology to examine driver’s license photos and catch identity thieves, but could this become a slippery slope? In fact, facial recognition has already been implemented in other ways. The Mirror reports that computer recognition successfully identified a rioter after the 2011 London riots, for example. This is actually pitiful compared to the human who recognized a hundred and ninety from the same footage, but technology advances rapidly.

2011, remember, was 5 years ago, and New York’s facial recognition technology was only introduced in 2010. It now has double the facial measurement points, which is the reason for the increase in solved cases. With the exponential nature of technology growth, facial recognition will likely become a ubiquitous part of our lives.

Surveillance: The Real Problem

But the London example highlights the fact that newer technology isn’t even necessary; the moral conundrum has already been posed. The Mirror also reported that, “By some estimates, as many as a million CCTV cameras are installed in London, making it the most surveilled place on the planet.” Since teams of humans are already successfully working with computers to recognize faces, the rising development of facial recognition technology is somewhat irrelevant.

The actual issue here is surveillance. The only reason facial recognition technology seems to be a more frightening aspect is because it cancels out one of the few rationalizations we have to feel comfortable under surveillance: anonymity through sheer volume of data. But again, due to human teams using computers to narrow down attributes like race and identifying marks and then sorting through the results to recognize faces, this is an illusion.

It will only become more of an illusion the more technology progresses, but we need to stop fooling ourselves now. This is the same rationality we use when we think about the very real surveillance of the NSA over our digital lives–there are so many people using the Internet, what are the odds they are watching me? But CCTV cameras are potentially even more restrictive, because they interfere with our right to travel freely. One does not need to be committing a crime to be embarrassed about visiting a certain store–and indeed privacy and surveillance are not just about crime.

That brings us back to our wrongfully-accused sci fi protagonist. There is a fine line between a rioter in London and a political dissident. Some would consider them to be one and the same thing. This is no big stretch of the imagination–several countries, even democratic ones, regularly jail and/or torture their citizens for speaking out against the government.

Yes, the technology has been used to catch criminals, but at the cost of imposing control on the rest of the population.

That also brings us to the issue of how and why facial recognition technology categorizes people. The same Mirror article reports that technology narrows down results based on racial categories: “The racial categories include Northern and Southern European, African or Caribbean, Middle Eastern, East Asian or South Asian.” Since western democracies often identify “Middle Eastern” with “terrorist,” we can imagine the implications for the surveillance of this racial category.

Surveillance as a Control Mechanism

What we are talking about here is the Panopticon–a phenomenon explained by Michel Foucault. It is an architectural phenomenon originally used as a prison. In this layout, the prison cells are stacked in an arc around a central guard tower. From the tower, the guard can see into any individual cell easily to monitor the inmates, but the inmates cannot see into the tower. Therefore, they do not know where the guard is looking, and have to assume they are being constantly watched. It is a mechanism of control.

Now the architecture isn’t necessary because we have the technology of CCTV cameras. The concept is exactly the same. On many cameras, there is a black dome so we know there is a camera but we don’t know which direction it is facing. We have to assume it could be watching us at any time.

The difference is that we have done nothing wrong and are not prison inmates. So, why is it that we are living in a “free” society divided into categories and controlled by surveillance as if we were criminals in a prison? Yes, the technology has been used to catch criminals, but at the cost of imposing control on the rest of the population, the vast majority of which has committed no crime.

So it turns out facial recognition technology isn’t really the issue. It’s the age-old moral quandary of surveillance. Facial recognition just intensifies the problem, because, increasingly, we can no longer be lost in the crowd.

  • Nic Barkdull is a PhD student in Social Research and Cultural Studies at NCTU in Taiwan.