There is a downside to surveillance, particularly if the data cameras capture falls into the wrong hands. (Photo: Chris McGrath/Getty Images)
Black Mirror’s “Nosedive” episode, coincidentally shot in Cape Town, was about a speculative future wherein social scores dictate one’s access to different realms of society. It might have been a bit too on the nose, if the intended pun can be pardoned. China has already implemented such a system, to terrifying effect, and it’s clearly visible in the responses of many Chinese public figures to the ongoing pro-democracy protests in Hong Kong.
There’s certainly something dystopic about this vision of our future: streets dotted with facial recognition cameras and demerits awarded for poor behaviour. Surveillance is a concept that many feel deeply uncomfortable about, as evidenced by the outrage that followed when Edward Snowden blew his whistle and informed the world that the American government was watching us while we went about our lives unaware.
According to the International Institute for Management Development’s Smart City Index, 83.7% of Capetonians would be comfortable with the use of facial-recognition cameras to combat crime. Facial recognition technology uses biometrics and machine learning in order to match what it sees with a database, and it sounds relatively straightforward as a policing measure.
But this is simply not the case. Studies on facial recognition software have found a considerable racial bias in many of these services, ranging from an inability to identify dark-skinned individuals to being a whitewashing portrait app. This algorithmic imbalance stems from a simple issue: that tech is an industry dominated by white men who have little to no perspective on how people of colour might interact with the products of their code. In South Africa, this bias could lead to a massive failing of facial recognition and has the potential to damage countless lives through its errors.
The solutions to these specific issues can be found in tech education. If the industry is intentionally and decisively opened up — if more South Africans of currently underrepresented races and genders learn to code and enter the industry — the software may well benefit from this broadened perspective, from both sociopolitical and purely technological points of view.
It’s important to remember that technology is not empiric and all-knowing. Instead, it is a product of the intellectual labour of people and as such is prone to mistakes, biases, and other issues that plague us as individuals. However, one must wonder if the simplest solution is to not adopt facial recognition at all. Despite the support evidenced in the Smart City Index, doubts are likely to appear should a plan for such a system be put in place.
In America, three cities have already banned the use of the technology and many digital rights surveys claim it should not exist at all. If one looks at its use in the suppression of riots in Hong Kong, it’s not hard to see why. Should we manage to create an equitable system for surveillance, what happens when it falls into the wrong hands? Information security is a growing priority for smart cities because it is a rapidly developing threat to them. Untold havoc could be caused by hackers, and access to a database of every face in a city would be a proverbial gold mine to such unscrupulous individuals.
These are the new dangers that the next generation must face, and one can only hope that their brightest minds are given the means with which they might map these uncharted territories.