There are some tasks we should never give up to machines. AHRI’s chairman, Peter Wilson CPHR, shares his thoughts.

The power of artificial intelligence (AI) has made significant in-roads into our working lives, and also the HR profession. Speaking with HR professionals over the last few years, I would have to say the jury is half in and half out as to whether this has been a good thing. 

Robot thinking and style have certainly impacted the most basic of HR practices. The best example of this would be recruitment.

Good recruiting practice now constitutes four activities: interviewing around the 360-degree dial; referee checks (preferably from a boss in the last workplace); credentials and skills checking; and psychometric testing. AI has had a big role to play in the last two items on this list.

Psychometric testing is now being widely used, whether as a screening device at the long list stage, as a fine-tuning mechanism with short-listing, or for final selection itself. Psychometric data raises privacy concerns, so it’s important in the job vacancy and advertisement stage to clarify how and when it is to be used with all potential candidates.

Don’t fake it

A majority of HR professionals have advised me that they are in favour of psychometrics, especially the more comprehensive versions such as the SHL, OPQ and CPI tools. The Myers Briggs Type Indicator (MBTI) test is a valuable way to encourage a discussion about personality types but it can be gamed. Participants who have done the MBTI a number of times can produce a profile aligned to how they would like to be seen, rather than their true underlying psychological type. I know this because I have done the MBTI test enough to see through how the results are calculated. My Myers Briggs type is called INTJ, but familiarity with this test has enabled me to identify the introverted/extroverted choice questions, and I can see ways to spruce up my answers. For example, I could look more extroverted if I wished.

 However, even Albert Einstein would have extreme difficulty gaming comprehensive psychometric test results like OPQ because they usually have a consistency score from 1-10 at the end. The psychometric algorithm will out you with a 1 or 2 consistency score out of 10 if you try to play games with answers across the 200 or so forced choice questions. It’s always better to answer honestly, and to recruit people who do this. 

The value of psychometrics is it reveals personal psychological preferences: how you like to work; how you are likely to respond under stress and pressure; how you relate to others in different environments; and whether you have a preference for creativity. These characteristics can be obfuscated or avoided in an interview. If that happens, the probability of a bad decision is increased. 

Beware of the risks

Professor Wayne Cascio, of University of Colorado in Denver, has estimated 20 per cent of external hires fail, so it’s important to avoid bad appointment decisions by investing in all four recruitment steps, and especially the unique insights available from psychometrics. 

AI is also having impacts in the big data stages of recruitment. Job boards and job aggregators like SEEK, Indeed, CareerBuilder, and Google for Jobs have hundreds of millions of CVs in their databases, and so are playing a huge role in making external hires more accessible and efficient. There are also applicant tracking systems (ATS) and automated assessment tools available for use. 

However, there is a critical issue with using these. AI is very good at combing through massive amounts of data and screening out critical information, fact and fiction, based on objective patterns and filters. It’s a good way to test whether job applications contain authentic and consistent data. But robots themselves are flawed interviewers for two reasons. 

First, when robots are used to screen candidates on their relationship-based interpersonal skills, it is inevitable that they will carry the bias of the person who coded the algorithm driving them.

Secondly, no matter how objective the coder is, it’s still an elusive task for a robot to ascertain correctly from its coding ‘how someone feels’ and the ‘nuancing of an interpersonal exchange and response’.

In our new world of man vs machine, I will always back the human touch to make that final assessment – notwithstanding our innate imperfections. We living breathing human beings won’t always get it right, but hey – it’s still a lot of fun trying. And it’s our natural right not to cede this ground to a bot. Don’t give up on that for any reason. 

This article was originally published in the October 2019 edition of HRM Magazine.

To learn more about the human side of recruiting, check out AHRI’s ‘Attracting and retaining talent’ short course for information on job design, interviewing methods and more.