Facial Recognition Is Not New (2019 Jun)

by Barry A. Liebling

Emerging advances in facial recognition have elicited a lot of discussion among technology enthusiasts, business owners, law enforcement professionals, and political pundits. Some private companies and government entities are using facial recognition systems to identify individuals automatically by comparing their image to a database of faces.

The technique can be used in retail to enhance the customer experience. Walk into a store, and a system will scan your face, recognize who you are, when you were last on the premises, and what products you have purchased. You might receive new offers and pertinent information that is tailored to your special profile. From the store’s perspective this can be a major advantage, since the system is keeping track of your visits and making it easier for the business to anticipate what is likely to catch your interest.

Facial recognition technology can also be used defensively. Some stores are using it to identify individuals who have been caught shoplifting or who have a history of attempting to return stolen merchandise. The automated system provides a warning to the establishment and alerts the management when to be vigilant. Law enforcement agencies are using the technology to scan faces in an attempt to find miscreants with a history of criminal activity. https://www.wsj.com/articles/big-brother-in-the-mall-11555128005

There are diametric ways of evaluating this new technology. Optimists see facial recognition as a welcome advancement that has great promise. It has the potential to make tasks that entail recognizing people easier, more accurate, quicker, and more economical.

But there is no shortage of pessimists. Negative reviews concern privacy issues and the possibility that the system will make errors. On the privacy front many critics fret that computer-assisted face recognition is reminiscent of George Orwell’s Big Brother. Go into a store, and your “right to privacy” will be violated. The store management will know who you are and when you have visited, even though you might not want to disclose that information.

Using the technology to detect and thwart criminal activity is even more worrisome to the pessimists. What if a face recognition system makes errors, and an innocent person is falsely accused of breaking the law? There is no question that even the best technology will have an error rate greater than zero.

Of course, to a modern Luddite the default solution to the “privacy problem” and inevitable system errors is to get government involved in passing new laws that regulate (and perhaps prohibit) the use of face recognition technology.

Think carefully about what face recognition technology is doing. Essentially it is making identification, classification, and judgments easier. Let’s start by considering what occurs when no technology is used to recognize people’s faces.

Think of an ordinary retail establishment. You walk into a store and because you have shopped there previously the manager recognizes you. You might be greeted and presented with offers that are calculated to be attractive to you. Is your privacy violated because your face is familiar to the manager? Do you have a “right to privacy” when you walk into the store, and should you be alarmed that your visage is associated with your name and shopping history? If it is acceptable for the store personnel to remember you without using any artifacts, what is the problem with enhanced technology? The smallest step in bringing technology into the store setting would be written notes. Do you object to the staff putting their experiences with you on paper? If ordinary unassisted memory is unobjectionable, what is the problem with recording observations in writing?

Now I will speculate on what a critic might say. The written notes might not be accurate. Even worse, the retail worker might deliberately write things that are not true which could be detrimental to the customer. And the critic has a valid point. Any technology can be used improperly or with ill will. Still, on balance, there are real advantages to having notes. If the person in the store is attentive and properly motivated the reliability and accuracy of notes should exceed what might be remembered with no technological enhancements. If notes did not do a better job than unaided memory, writing things down would be recognized (eventually) as a waste of time.

What new laws should be passed when retail establishments use written journals instead of relying on what they happen to remember in recognizing and dealing with customers? None! If there are penalties for making a mistake (or slandering) someone without using notes, putting thoughts on paper should change nothing. Individuals are responsible for making good judgments (regardless of what artifacts they use), and they should always be prudent.

Note that a similar line of reasoning applies to identifying people defensively. Police that use notes are on balance more accurate than police who do not. Law enforcement errors are bad and when they occur they need to be corrected – whether or not notes (or any technological methods) are used.

It should be apparent that if it is better to use notes than to rely exclusively on memory, it is probably even better to add photographic images. And photographic images (which have been used advantageously for more than one hundred years) can be made even more accurate with face recognition systems.

Again, let us return to the critic who wants to reject the use of face recognition devices. The adversary of the new technology might say that the new systems are error-prone and are not more reliable or valid than conventional ways of identifying people. And again, the critic might have a valid comment. Automated face recognition should not be used unless it can be demonstrated that it is genuinely effective (in conjunction with properly motivated and trained human managers). And, not incidently, that is why some private retail establishments and some law enforcement agencies are employing it – and some are not. Before adopting a new device or system the intelligent user checks to make certain it is working properly.

Here is the bottom line. Face recognition is not new. Humans have always been doing it. Advances in high technology are a net positive. Just use them responsibly.

*** See other entries at AlertMindPublishing.com in “Monthly Columns.” ***

Comments are closed.