Recruitment Grapevine | Executive Grapevine International Ltd

Faceless Hiring

The dark side to using AI to screen out candidates
Faceless Hiring

Technological innovation over the last decade has been vast. From driverless cars to Artificial Intelligence (AI) supporting disease diagnosis alongside home assistants dimming lights at our vocal commands, these feats are no longer confined to a visionary episode of The Simpsons – they are our reality. Whilst advancements in AI, algorithmic and machine learning have arguably enriched the human experience, our acquiescence of digital into both our personal and professional lives hasn’t been met with much resistance. Now, however, there is concern over a future deep-rooted in digital, particularly regarding the threat of automation on human labour and trust in algorithmic decisions.


The latter is an issue recently highlighted in the hiring space. Securing a job predominately takes place in online, with Applicant Tracking Systems (ATS), automated screening tools and online assessments wiping out the need for handwritten CVs and cover letter – and in some cases, face-to-face interviews.

Although these tools have enabled a much faster, efficient experience, common complaints for candidates, centre around the process being both frustrating and faceless. A recent Guardian report followed the journey of jobseeker Deborah Caldeira, who holds a master’s degree from the London School of Economics and has had 86 unsuccessful job applications in two years. She said tech-based hiring platforms can be disillusioning, as it’s hard to know: “exactly what the robot is looking for. It makes us feel that we’re not worthwhile, as the company couldn’t even assign a person for a few minutes.”


Another graduate jobseeker, Peter Lane, told the BBC that when he was job-hunting, he secured around 15 interviews – but never met any potential employers. "They were all video-based screening interviews," Peter said. "There was no way to tell if I'd impressed them with my answers or experience as there was no human interaction." Another issue included the automated rejection letters. "Only 10% of potential employers have given me detailed feedback," he says. "As jobseekers, we need to know where and how we can improve - whether that's with our CVs, job experience or even personality." Twitter has also acted as a cathartic outlet for frustrated jobseekers.

Learning to speak ‘robot’

This is a problem frequently witnessed by Victoria McLean, founder and CEO of City CV.co.uk, a CV advisory firm. Citing research that suggests 80% of CVs are rejected by ATS filters within 11 seconds, McLean says it’s “the most frustrating - and sometimes heart-breaking - aspect of job hunting from a candidate’s point of view. They’re falling at the very first hurdle.”

“Just a few weeks ago I took a call from someone looking to take the next step in their marketing career,” McLean explains. “She had a business degree and seven years of relevant experience. But she’d been applying for jobs for six months and wasn’t getting any interviews.” But, when McLean and her team examined her CV, it became apparent that she just hadn’t learned to ‘speak robot.’ “We undertook robust keyword research and re-wrote her CV, placing the right keywords in the right places, in addition to giving the CV impact,” she said. “When she posted her new CV on jobs boards, she started getting interviews within 24 hours. It really was that quick.”

 

Frustrating, faceless but fair?

Clearly, becoming fluent in ‘robot’ is one way to appease these digital ‘recruiters’. And, according to a Hays Australia study, candidates are already adapting with these nuances in mind, with 81% saying that have or plan to adapt their CVs and online profiles for initial screening by an algorithm. 27% say they are already doing so. However, even when candidates do get past the algorithm, McLean isn’t convinced that machines can choose the best talent. “This is particularly the case with candidates who haven’t followed a conventional career path or have taken a career break,” she says. “Yet, these are the very people who can bring extra skills, a different perspective and more diversity of thought to the organisation.”

It's in this word, ‘diversity’, that the debate for and against algorithms in recruitment hinges on. Despite technology promising to assist in eliminating bias – by assessing candidates on a meritocratic basis - it can also perpetuate it. “One limitation is that algorithms feed off historical data, perpetuating built in biases,” Amit Mohindra, former Global Head of HR Analytics at Apple explains. For example, if an organisation uses a sample of their top performers to determine the qualities of a successful candidate, if the sample is from a homogenous group, the algorithm might be unable to assess those who don’t fall into this category.

“One in four of us in the UK now works flexibly and we’re seeing a new generation of individuals enjoying the freedom and benefits that being a contractor can offer”

 

The same goes for video interviewing. Acting as a halfway point that brings more personality into the process, whilst simultaneously allowing recruiters to sift through applicants quickly, video-interviewing platforms like HireVue are already being endorsed by city firms such as Goldman Sachs and Unilever. The software works by scanning candidates for emotion and expressions, such as blinks, smiles and frowns, and can “provide tens of thousands of usable data points” Gema Ruiz de Huydobro, Ph.D., IO Psychology Consultant at HireVue, tells us.

However, relying on these ‘data points’ presents its own challenges “What tends to happen is that employers analyse the vast amounts of data they have on their current employees,” McLean explains. “They then use this to build a profile of the ‘ideal candidate’. The risk is that implicit or unconscious bias has been programmed into the AI and you end up hiring people who are just like you and your current team.” Yet, as Ruiz de Huydobro argues, Hirevue, just like other firms well aware of this potential for bias, are doing all they can to not prejudice certain candidates. ““From the beginning,” the psychology consultants outlines, “HireVue has developed a rigorous testing and validation process which is primarily focused on finding and eliminating any factors that show biased results for protected classes.  This process carefully tests for any bias that may be present in input or output data before, during, and after development of the algorithm, and thorough testing continues to be performed as part of an ongoing process of prevention.”  This kind of best practise is what all recruiters need to gird their processes with otherwise AI-enabled hiring will not be the panacea it could be.

And, there is another way. Like Mohindra, Jan Mueller, Global Vice President of Marquee Accounts at Korn Ferry, advises organisations to thoroughly audit their data instead of letting a machine dictate. McLean concurs, adding recruiters should, where they can, be transparent. “Candidates don’t know why they’ve been rejected,” she argues. “It’s not like a traditional face-to-face interview where you can ask for feedback and work on improving your interview technique and skills.”


The convenience of digital

Whilst digital recruitment might feel faceless, Jonny Luk, Author of TheGradJobGame told us that when job-hunting, he enjoyed being able to submit videos or links to portfolios, as it enables candidates to portray themselves creatively - meaning applicant data needn’t also be faceless. Although he acknowledged that the pre-interview stage can be impersonal, with job boards ease of use and ‘one-click’ apply systems encouraging mass application, it’s justified. “The benefit of having the initial filtering so brutal is that the later stages tend to be more personal,” Luk explains.

Ruiz de Huydobro is also positive that technology is enhancing the candidate experience. “One of the common misconceptions is that this approach makes the recruitment experience less personal – but when you compare it to previous methods, this claim becomes unjustified,” she explains. “Typically, companies who have implemented AI in their recruitment process are large organisations with a high volume of hires. For decades, these organisations have used traditional assessments to provide an objective, fair and science-based way to match candidates to the competencies required. However, the experience was often both poor and time-consuming. Companies were not able to consistently hire the best people; instead they were hiring the best people who were also willing to endure their hiring processes.”

“A candidate can then spend less than 30 minutes completing an AI-driven pre-hire assessment on their smartphone”


Now, Ruiz says, candidates don’t need to spend three hours completing cumbersome applications. “A candidate can then spend less than 30 minutes completing an AI-driven pre-hire assessment on their smartphone,” she explains. Fortunately, this is matching candidate expectations. According to a separate Hays report, 71% of applicants would lose interest in a role if the online process took over a quarter of an hour.

However, despite the risks of bias being built into AI, if recruiters are cognizant of how the technology they use might overlook talent and aren’t afraid to challenge it, the benefits are clear – for jobseekers too. Today’s tools are a “step towards the right direction” adds Luk. “We just need to ensure that technology continues to make it easier for both recruiters and jobseekers to find the job that fits them.” What’s clear is that algorithms – no matter how seamless or bias-free – cannot replace the value of human interaction nor judgement when it comes to employment. After all, you can’t ‘shake on it’ online.