Can You Hire Using Artificial Intelligence?


The use of technology in acquiring talent is a well-established conversation.  The use of artificial intelligence to influence the way in which applicants are screened for job positions, however, is a concept that appears only now to be gaining some traction.

This is a bold move, and one which could profoundly change the existing paradigms of talent acquisition.  Essentially we are talking about removing the need for a human to be involved in applicant screening at all, which can save organisations vast amounts of time and money.

However, this is not as revolutionary as it sounds.  A large proportion of big organisations already use online screening tools such as Situational Judgement Tests (SJTs), ability and personality inventories and assessments of technical capability (a&dc have been producing these for our clients for years).  If you ensure they are measuring the right indicators of performance for the role you’re hiring (by doing the role research scientifically), then you have the ingredients for a pain-free and effective acquisition process without needing hours and hours of human input.

Why?  Because all of these techniques have automated scoring algorithms that can easily be integrated with applicant tracking systems to progress people to the next stage of a sifting process.  Let’s be clear though; this is not artificial intelligence.  This is simply a computer being a computer.  This includes the use of software to scan CVs / resumes for keywords like a search engine, which unfortunately is a rather bad predictor of future performance (with research showing correlations barely more effective than methods like graphology and astrology).

So what do we really mean by the term ‘artificial intelligence’?  Well, it refers to a computer system that can perform tasks typically associated with human intelligence.  To name a few, these abilities include speech recognition, visual perception and decision-making based on a sensory input.  Some would say it includes the ability to learn.  Psychometric tests have the capability to be ‘adaptive’ (directing more or less difficult questions to test-takers based on how well they have answered previous questions), and we can use ‘self-learning’ algorithms that can interpret data sets for us.  But even these fall short of properly demonstrating intelligences similar to human abilities.

Here’s the interesting story; some organisations are attempting to use software that can analyse video interviews by recognising the body language, vocal pitch and pacing of an interviewee’s responses, and mark them against pre-defined criteria.  So what are the advantages?  Humans have biases, while computers are 100% objective and more consistent.  By also reducing the people-hours needed, you’re creating an exciting cost-saving and a theoretically higher rate of accuracy in predicting performance.  The most significant thought, as these techniques get more and more sophisticated, is that this could remove the need for humans to be involved in any stage of recruitment.

The technique, as with most recruitment methods, does have some issues that need to be overcome.  After all, it is still at the mercy of the humans responsible for programming ‘what good looks like’ (as computers cannot learn this for themselves yet, it means they do not have proper artificial intelligence).  If unconscious bias influences the marking criteria even slightly, it could cause big repercussions in how accurately the computer judges a candidate’s performance.  This immediately undermines one of the major advantages of replacing assessors with computers.

Undoubtedly, innovation in technology continues to astound us, but we need to be cautious in what gets branded as ‘artificial intelligence’.  The answer to the title of this blog post is therefore currently ‘no’, but this could change in the future.

A big question will remain if this does change; what happens to the candidate experience?  People need to feel comfortable to behave genuinely, and genuine behaviour is necessary in order to make accurate observations and hiring decisions.  As a result, here is my takeaway message: technologies should be used in hiring, but their usage should clearly be balanced with a human touch.

Author: Jordon Jones