It’s almost impossible to talk about recruitment these days without seeing or hearing something about AI or machine learning. That’s not a bad thing of course – we live in exciting times technologically. There’s already a plethora of advanced recruitment tech solutions out there, claiming they can do everything from interviewing candidates better than a human to sifting through CVs to pick top applicants. We hear a lot about the advantages, but are we talking enough about the potential downsides? Here, our partners JobLookUP, share their thoughts on the 3 major problems with AI in recruitment.
Predictions that AI could replace human recruitment specialists, or that HR could end up a fully automated process are often spoken of as the potential costs of innovation. While there would be a certain amusing irony in seeing an HR department run entirely by machines, the human element of recruitment and HR practice is only likely to become more important, not less. Understanding this ‘human factor’ (which includes soft skills like emotional intelligence and decision making) in recruitment is also the key to highlighting the potential risks an increase in AI and automation carry with them.
Here are three possible problem areas that could cause major issues for recruiters and jobseekers alike.
Automated Recruitment can be Exploited
We’re all getting more tech-savvy Thanks to the internet, as soon as a new piece of tech or software is out there, it’s often only a matter of hours before someone has worked out how to hack, circumvent, or modify it. New recruitment tech is no different, and ATS systems are the perfect example.
Research conducted by Tribe Pad, a software provider, found that 88% of candidates who are aware of applicant tracking systems have attempted to optimise their CV. That’s normal enough, but one in four those aware of ATS have also tried to game assessment systems. The study also found that:
- 57% of candidates go for quantity over quality when applying for jobs
- 42% worry that tech might screen people incorrectly
This combination of distrust in recruitment technology, and the desire to get ahead in the hiring process however they can, means jobseekers have started using a variety of measures to try and manipulate the recruitment process in their favour. Some tactics include:
- Searching online to see how other people got hired at the same company, or for a similar role.
- Using keywords or buzzwords that they know will get their CV search hits from an ATS or automated system.
- Exaggerating or lying about previous experience on a CV.
- Cheating on skills or psychometric tests.
While the percentages of surveyed candidates using these methods wasn’t that high, it is significant, and gives us one example of why recruiters can’t rely on technology alone to choose applicants.
It could be argued that, in a way, nothing has changed. Exaggerating on a CV isn’t a new trick, and neither is looking up what other candidates might have done to get hired. The problem is, AI might not be able to spot these issues. An experienced recruiter, on the other hand, could see the potential warning signs.
AI Bias and Transparency Could be a Problem
Around 42% of respondents to TribePad’s research thought that technology dehumanises the recruitment process, and that recruitment decisions should ultimately be made by people, not programs. A smaller number (21%) highlighted what could be a much greater problem – that software and AI are programmed by humans, and are therefore likely to inherit inherent biases as a result. As it turns out, this unease isn’t entirely misplaced.
Firstly, there are the ethical challenges of handling such detailed personal information. In a world where data is now almost the most valuable commodity (set to be worth around €111 billion in 2020), how we approach the privacy issues that could arise as a result of AI or automated systems using and sharing CV information is something that needs to be carefully considered.
Then there’s the question as to whether the qualities that make a good employee are even able to be parsed by AI at all. Many experts have their doubts. Meredith Whittaker, the co-founder of the AI Now Institute, suggested claims that AI can spot good employees through facial movements, mannerisms, and tone of voice are nothing more than pseudoscience. She’s not alone in making such criticisms.
Another area of concern is transferring bias directly into automated systems and AI. Can a facial recognition AI accurately read the expressions of all people, regardless of age, gender, and race? Can an automated recruitment system really construct an accurate picture of an applicant and their personality from a CV?
Unless these systems have been programmed with an almost uncountable number of variables – currently an extremely difficult task – then they will operate within the limits, and biases, of their programmed parameters. Who takes responsibility for discrimination and bias in these systems? The HR department? The recruitment agency? The software provider? No-one?
Jobseekers Prefer Human Recruiters
The final, and arguably most fundamental issue, is that jobseekers simply don’t want to go through an entirely automated hiring process. And who can blame them? We only need to look at an automated switchboard to see how infuriating a human free recruitment process could become if we’re not careful. In the TribePad survey, 75% of jobseekers said they were also concerned that their CV wouldn’t make it past screening AI or software, while 70% said they thought interviews should be conducted by an actual person.
This might all paint a fairly negative picture, but that’s not to say we shouldn’t use new and innovative tools to increase efficiency, reduce cost, and improve processes. Of course we should. Jobseekers know this too, and despite having many reservations about up and coming technology, most use online job boards by a clear margin (56%) to find their next role these days.
There’s plenty of positive feedback for video interviewing, skills tests, chatbots, and automatic CV reading software as well, so it’s clear that recruiters don’t need to be taking any extreme steps back into the dark ages.
A Tool, Not a Necessity
In the end, like many other industries, there’s no doubt that automation and AI will affect the way recruitment works. That doesn’t automatically mean removing people from the equation entirely. People are going to be more important than ever, in fact – it will be those exact qualities that make us human that means we’ll be an essential part of any process, especially in a sector as social as recruitment. Human recruiters aren’t going anywhere.
New technology and its potential is always exciting. We just need to ensure we measure the impacts, advantages, and disadvantages carefully, and objectively.
Author biography:
Simon Marshall, Head of Content and Communications at JobLookup.
Simon is a technology enthusiast with an interest in how the application of science, philosophy, ethics, and psychology can benefit us in both the workplace and wider society.
