Regulating the use of facial recognition technology

It’s not every day that you hear a company that manufactures a particular product ask that the product be regulated by the government; certainly very altruistic and noble, but not commonplace. Except you are Microsoft, which has often times showed itself as a socially responsible company; certainly an example for the likes of Facebook and […]

It’s not every day that you hear a company that manufactures a particular product ask that the product be regulated by the government; certainly very altruistic and noble, but not commonplace. Except you are Microsoft, which has often times showed itself as a socially responsible company; certainly an example for the likes of Facebook and Google to follow. Recognizing the potential abuse in the use of facial recognition technology (FRT), Microsoft is asking the U.S. Congress to study facial recognition and oversee its use.

The motivation for calling for regulation derives from the fact that the FRT has grown to be so powerful; and yet very easy to abuse. After all we hear about abuse in several parts of the not-so-free world, as well as in the U.S. The FR can be used to identify people in photos, video feeds, and of course in person, without their knowledge or permission. Certainly, the technology is invaluable for identifying criminals, but it also can easily and quite trivially be used to violate people’s privacy. For example, the FRT enables mass surveillance, and could therefore control people’s willingness to freely participate in activities where they would rather prefer to be anonymous.

In the 11 June 2018 article in this column in Daily Trust, I wrote on the use of high tech to secure developing countries, with a particular reference to Nigeria. I pointed out that, besides the use of CCTV, other high tech platforms may be needed to help with the solution of the country’s security problems. I also suggested that the idea of massive deployment of facial recognition technology (FRT) should not be off the table. Of course, I am not in support, or an advocate, of the violation of human rights. However, I do believe that tough situations require tough measures. In terms of security, I suggested that China might actually be a model for Nigeria. I wrote this: “With a population of over 1.3 billion people, I don’t think China has more than a handful of gun violence in an entire year. Yet, Nigeria, with barely one quarter the population of China, has a crime rate that is orders of magnitude higher than that in China. From the ground, it is really not high technology that has kept China safe, but these days, it seems the need has arisen to maintain the safety status quo by massively deploying FRT.” 

I further stated in the referenced article that “The FRT can be deployed in banks, airports, garages, markets, office buildings, and so on. With the massive deployment of FRT in China, coupled with the usually lower costs of gadgets from there when compared with the West, it is not inconceivable that Nigeria would head east for its FRT solutions.”

 However, I am not oblivious to the potential misuse of FRT – by governments, organizations, or individuals. Furthermore, we cannot pretend to be unaware of the problems with FRT. As pointed out in previous articles in this column in Daily Trust, the accuracy of FRT and virtually all artificial intelligence (machine learning) tools depends on the data on which they are trained. Biases of the developers eventually show up in the tools they develop. For example, because current FRT tools have focused on white (and perhaps yellow) people, the most popular FRT apps, such as those developed by Microsoft, IBM, and Megvii of China are grossly inaccurate for identifying the faces of people with dark skin. This means that FRT tools developed in China for the Nigerian market has to be trained accordingly. Unfortunately, unlike their counterparts in the West, it does not appear that many vendors of Chinese FRT tools have addressed the crucial issue of bias.

The need for regulating the use FRT seems obvious from the foregoing. In April of this year, privacy groups in the U.S. filed a complaint with the Federal Trade Commission alleging that Facebook had turned on new FRT services without obtaining the permission of users. Obviously, Facebook is not a government regulatory agent; so why is the company collecting folks’ facial data without permission? This is important since this company has demonstrated its ineffectiveness in handling users’ personal data. In May, a few civil rights groups in the U.S. asked Amazon to stop selling its face-matching service, Rekognition, to law enforcement agencies. (Apparently, anyone can license Rekognition!)

Mr. Bradford Smith, Microsoft’s president, suggests that “governments around the world should examine law enforcement and commercial uses of facial recognition.” The European Union (EU) is way ahead of the curve here; thanks to its tough and proactive new data protection law, which prohibits companies from gathering the biometric data needed for facial recognition without first obtaining users’ specific consent. The EU’s General Data Protection Regulation (GDPR) that went into effect on 24 May 2018, with enforcement starting the following day, provides for this defense, among others.

Apparently, Microsoft and Civil Rights groups are appealing to governments for similar controls in the U.S. and other places.