We See You!
While facial recognition technology’s application is growing exponentially, so too are the concerns of privacy, accuracy and liability.
Person of Interest is a television show that ran from 2011 to 2016 which chronicles the adventures of an ex-assassin and a tech geek who collectively prevent crimes with the use of a super-sophisticated video surveillance computer. At the heart of the show is surreal facial recognition technology. While aspects of the show were clearly science fiction (the technology that used the facial recognition was called the “Machine” and seemed both omnipotent and omniscient), nary two years after the end of the show, facial recognition headlines in the news have become ubiquitous.
Facial recognition technology is all around us. Microsoft’s Surface tablet and Apple’s IPhone X, as well as many other consumer electronics, now use facial recognition technology for login. Toronto city council just approved the doubling of the number of surveillance cameras throughout the city, at the request of the police following the brutal shootings on Danforth Ave. Facial recognition technology is all around us even when we can’t see it at work. Several Canadian malls use cameras and face recognition technology to guide shoppers and enhance the retail experience. Amazon’s shareholders and employees recently pleaded with Amazon not to license its Rekognition software for clandestine law enforcement use. And this was just a sampling of the headlines from the last month alone! Facial recognition is not only here, but it is here to stay.
Facial recognition technology is far from new and has been around in a rudimentary form for decades. It is only within the last few years that facial recognition technology has made exponential leaps in practical application. This is, in part, because of improvements in camera technology. More importantly, the algorithms used to differentiate images have become that much more sophisticated and smarter. AIH Technology Inc., a Toronto-based company specializing in facial recognition technology originally developed by its China division, notes that, unlike your grandfather’s facial recognition software, the modern technology is optimized to recognize faces of all ethnicities, in motion, in vehicles, in disguises, and in extreme low-light conditions, all in near-real-time with 99 per cent plus detection accuracy rates.
Of course, even non-lawyers realize that few new technologies have attracted more legal attention than this modern facial recognition technology. There is a whole tension between privacy and security and where that line should be drawn. Currently, there is very little legislation or case law directly on point, even though the technology is already widely in use. This isn’t the first time that laws have significantly lagged the development and widespread use of technology. Only recently have governments begun reacting to the almost exponential growth of this technology. For example, while it does not appear that the shopping centre operators did anything at all wrong by installing cameras in their malls and tracking customers through those cameras, nonetheless both the federal and provincial privacy commissioners have opened an investigation into the use of facial recognition technology in large public shopping centres.
While it is still clearly the “wild west” in terms of ethics and legality, things are rapidly evolving along the legal frontier and we anticipate comprehensive rules to formulate soon, either through legislation or case law. While a detailed legal framework will eventually emerge, we do not interpret this as meaning, in any way, that the growth of the technology will be materially curtailed. If anything, we feel that the rules that will emerge will focus on the protection and potential abuse of the resulting data (rather than the surveillance per se) and will impose liability on false identifications. Most readers are familiar with the recent experiment where Amazon’s Rekognition software falsely identified 28 members of the U.S. Congress as known criminals. We can see the law not being tolerant of any actions based on that kind of inaccuracy, but any such laws, rather than limiting the use of facial recognition software, will simply give the competitive edge to technology providers with better, stronger and faster technologies.
But don’t take our word for it. If capital markets are any predictors of the future, then the money seems to agree with us in positing that facial recognition technology is here to stay. The New York Times recently ran an article discussing emerging Chinese companies leading the charge in advanced facial recognition technology (the best facial recognition technology all seems to be coming out of China). In that article, SenseTime, an artificial intelligence company based in Beijing and specializing in facial recognition, recently raised over US$1.6 billion, giving it a market capitalization in the US$4-billion range. Similarly, Megvii, the developer of a popular Chinese facial recognition software system called Face++, raised almost US$600 million in public funding recently. But perhaps more telling than these fundraising stories is the identities of the investors putting their hard-earned capital into this nascent technology. Investors in these Chinese facial recognition technology companies have included Chinese heavyweights like Alibaba, Bank of China, Foxconn and Ant Financial, as well as American financial institutions like Qualcomm, Silver Lake, Tiger Global Management, and Fidelity International.
One of the reasons that we do not anticipate legislation eliminating the use of facial recognition technology is because it simply is that good. We feel that the term “paradigm shift” is overused, but facial recognition technology really represents a paradigm shift in building maintenance, retail strategy, every imaginable security application, and so much more. While the “Machine” from Person of Interest is still perhaps science fiction, modern facial recognition technology is not that far behind, and it is here to stay.