Who's afraid of automation? - Machine learning and AI for recruitment
Automation is everywhere. It’s the new normal. As consumers, the world around us has been transformed by automation. Google Maps, same-day delivery, taxi apps, banking, the list goes on. In fact, it’s now considered a poor experience if we need to use our phones to speak to someone.
Machine learning and artificial intelligence are in our everyday lives making things easier, quicker and more convenient. So, what’s so wrong with using it in recruitment?
The “hollowing out of human interaction”, “artificial barriers” and concerns that automation “draws on data that’s often been shaped by inequality” according to a recent piece in The Guardian. “The grim reality of job hunting in the age of AI” really is grim according to this critique by Stephen Buranyi.
A bewildering, impenetrable selection process in which alienated candidates receive no feedback is a poor recruitment experience indeed. But any recruitment process, whether automated or not, can be done badly, or well.
I’m surprised that recruitment automation has engendered such a pessimistic assessment. When in fact, I’d argue that:
"Automation is what candidates expect, and what they want."
Consider the interactions we have with companies as consumers. Automation helps them provide a more personalised and more efficient experience for us. Likewise, candidates expect instant access, instant decision-making, self-service options, a fast and slick online experience.
Candidates don’t want to wait for decisions. And If they’re not successful, they want instant feedback in return for the time they’ve invested. And they want all of that to be accessible from their mobile phone.
“The standardised CV format allowed jobseekers to be evaluated by multiple firms with a single approach” laments Buranyi. Yes, often to their disadvantage. Many of our clients champion CV free hiring. Why should a chef with English as a second language be judged on their ability to fill two A4 pages and write a personal summary?
Surely by the time candidates get in front of hiring managers we should already know they’re able to do the job? At the apply, screen, assess stages, automation matches behaviours, values, skills and potential to a role in a way the generic CV can’t, meaning that the very best candidates are sent through to face-to-face or assessment centre.
While it’s argued that “jobseekers are forced to prepare for whatever format the company has chosen” the selection process is actually simulating real-life job demands - perhaps attitude to risk, ability to prioritise or work under pressure. The end result being that more people end up in the right role.
And “what’s the cost to workforce diversity?” It’s true that machines can learn bad habits but algorithms are not just our opinions embedded in code. Any kind of bias can be stripped out of an algorithm. The same can’t be said for humans whose decisions can be affected by unconscious bias, tiredness, stress and even hunger.
Automation can systematically address reviewer inconsistency, conscious, and unconscious bias, so decision-making becomes consistent and fair and diversity is improved. Conversely, we know that when:
multiple human reviewers score a single candidate, they disagree 95% of the time, and strongly disagree 49% of the time.
The reasons for adopting automation are compelling. A more engaging candidate experience, better decision-making, better workforce diversity and more people in the right jobs. It’s far from grim. Automation is not something to be afraid of. But like any other business process it needs to be done well.
If a company is succeeding in using automation, candidates will find that recruiters actually have more time for conversation and exchange, as automation frees them up from a huge volume of manual tasks.
With automation, the experience for candidates will be more human when it matters.
If you’re interested in how automation can be applied to improve the recruitment process, you can download our new guide for recruiters here.