Win the Talent War with Affective Computing
This post was originally published on the EnableX blog page
Job interviews are dreaded by all! It’s not just the potential candidate who feels the trepidation and stress of the experience, but even the interviewers find themselves nervous and unsure about the whole process. If you ask managers whether they consider themselves good interviewers, many will not affirm confidently. Interviewing is something that we think we know all about in theory, while in practise it may not be the case.
Uncovering the Essence of Correct Judgement
At its very elemental level, an interview involves an exchange of information. Usually, it’s the candidate who provides most of the details, while the interviewer records and processes it and arrives at a decision. The information can be factual such as age, gender, location, etc., or it can be empirical, like the candidate’s views on some issue, their ability to reason and more.
Evaluating these empirical data points is where the real challenge lies. If I want to select a candidate with the traits of “confidence” and “assertiveness” how do I go about it? I fall back on my understanding of these traits and try to gauge whether the candidate’s behaviour matches my understanding. The basic premise of my understanding of the traits creates the first point of failure. Maybe what I consider confidence, is not really what it is. Being humorous in an interview could be considered a sign of confidence, but it could also mean that the candidate was speaking from a prepared script. The end judgement may or may not be correct.
Hiring the Right Match with Emotion AI
In the century we are in, technology is touted as the solution for all problems, and this problem is no different. Affective Computing, aka Emotion AI, is now providing solutions to better evaluate these aspects of a candidate. But what is it based on? Many luminaries, such as Paul Ekman and James Russel, proposed credible theories to establish that human emotional states can be identified through the measurement and analysis of distinct physical responses, and these have become the cornerstone of Emotion AI technology.
Certain communication and AI API providers use the models of Ekman, Russel et al to create a mechanism to read a person’s facial expressions and break them down to categorise their emotional state. Basis their facial expressions, the presence and the respective intensities of seven core emotions — anger, disgust, fear, happiness, sadness, and surprise are determined. Along with this, the dimensional model is able to map these measurements to over 90 separate moods/states, which gives detailed information about the candidate in real-time.
Learning the Mechanism behind Emotion AI
Think of it as an emotion detector connected to the candidate. If I want to assess, let’s say the positive traits of the candidate, I can set up indicators that show the level of confidence, honesty, nervousness and attention the candidate is displaying. So if I see a candidate’s confidence and honesty level dip when they are talking about their impressive achievement, it’s a fair guess that they could be lying.
Instead of depending entirely on their personal evaluation of the candidate’s characteristics, HR managers and recruiters can now get the AI to validate/invalidate their opinion. During an online interview, a hiring manager, for example, might feel that the candidate is responding truthfully to a question, whereas the AI may indicate a high probability of lying. Seeing the emotional state of the candidate when answering specifically designed questions can help in gaining a deeper insight into their mental makeup, allowing better validation of non-quantifiable data points.
The ability to use this technology is becoming easier by the day. Integration into video conferencing platforms, including self-recorded video interviews, can help evaluators gain deep insights about the candidate, without them spending time doing a long, personal interview. Various attributes can be clubbed into consolidated metrics which can be set as thresholds for further evaluation, thus saving time, effort and cost in the entire recruitment process.
And given the current technologies, the entire process is done in real-time, without storing any personal or private data on the candidate, ensuring that it will be no different than meeting face-to-face. Is this the future of effective recruiting? I certainly feel that there is a lot that Emotion AI can do to change the very nature and output of the recruiting function.
To try it out, simply sign up here for free