The Surprising Tech of Facial Analysis

Article main image
Dec 17, 2018

Seemingly out of nowhere computers have learned to read human emotion from our faces. We can expect that any weaknesses this technology has now, in terms of accuracy or cost, will quickly erode as the technology advances. Marketing is interested in this technology to learn more about customers; HR is interested in using this tech for screening candidates.

In a nutshell, the computer does just what you or I would do: It looks at a person’s expression and categorizes it as happy, sad, surprised, fearful, angry, delighted or some other emotion. It can also track what a person is looking at. For example, if the customer shows delight when you show them your new product, but the scan shows they were actually looking at the donuts at the back of the room, then you’ll have to draw your conclusion.

When would HR want automated measures of emotion and attention? In the future, HR might be interested in how people are reacting to a training program;Hhow candidates are reacting to a job offer, or whether workers who are operating dangerous equipment look tired. At the moment, the “ready now” application is in recruiting.

The start-up Knockri assesses foundational competencies such as empathy, collaboration, or growth mindset based on how a candidate responds to a behavioral question. A candidate’s facial expressions, as well as the tone of voice and an automated analysis of the content of their answer, provide clues as to how high a candidate scores on the targeted competency. It adds not just automation to the screening process, it can actually reduce human bias.

There are many other vendors working in this space, such as the Emotion Research Lab based in Spain and the well-known recruiting tech firm HireVue.

What is interesting?

  • A technology that was almost unimaginable a few years ago can quickly become universal and cheap. This will keep happening so buckle up.
  • Humans react to technologies that invade our privacy with shock and horror — for about two weeks. After that we get used to Gmail reading our emails, LinkedIn suggesting who we should be friends with, and our smartphone tracking our every move.
  • We don’t need to become experts in this technology because vendors will package it up in a way where it just does what we want simply and seamlessly.
  • We can speculate on where this technology will prove to be most useful, but we’ll probably be wrong.

What is really important?

  • Even though we don’t need to be experts in the technology to use it, we do need to be experts in assessing if it is delivering what it says it will, and whether it complies with ethical and regulatory matters.
  • Any big long-term investments we make in a process may be rendered obsolete by unforeseen technology. For example, you may invest in some complex assessment methodology only to find it upended by a new and better way of doing things.
  • It may not be one killer technology that revolutions HR but a whole suite of new tools that together lead to a process that is significantly more effective than the competition’s. Imagine a recruiting process that uses superior sourcing technologies, deploys more effective job ads, creates a great experience for candidates, uses a suite of accurate assessment technologies, and constantly gathers data on the process so that the recruiting R&D team can ensure the great process continually get better. That’s probably where the big win will be.