Should You Use AI to Interview Job Candidates?
This piece originally appeared on the Workato blog.
People in tech are constantly talking about its potential to change every aspect of our lives. From the way we work to the way we eat, shop, and even sleep, it seems like technology is constantly rewriting our every activity.
Some of the most exciting changes are the ones that stem from the application of cognitive technologies like artificial intelligence (AI) and machine learning (ML). Aided by these smart tools, we can not only live differently but think differently: better and faster than we could possibly think on our own.
But what happens when we trust judgements and experiences that used to be fundamentally human to machines? What if computers had to make decisions about people’s value and potential? Would that be ethical — or desirable?
Today, companies are wrestling with this question as they look for better ways to find and attract talent. One of the most recent trends — using AI to conduct initial screening interviews — has provoked nothing short of full-blown controversy. Employers say it’s a much-needed cost-cutting tool, but job-seekers are skeptical. Are their concerns justified?
Winning the War for Talent
To understand why companies would consider using AI to interview candidates to begin with, you first have to understand two crucial facts about the job market:
Today, the unemployment rate in the US has hit an historic low. At 3.7%, joblessness is at its lowest point in 50 years.
At the same time, companies are looking for new talent more than ever. In 2017, for example, LinkedIn projected that hiring volume would increase by 58% in the US.
Because there are fewer job-seekers and more jobs that need filling, talent now comes at a premium. Companies need to hunt harder and longer to find the right candidate, and they need to make themselves more attractive in order to woo prospective employees away from other opportunities.
All of this adds up, both in terms of actual cost — offering great compensation and cool perks isn’t cheap — and in terms of opportunity costs, like the amount of time HR staff spend screening resumes. According to Glassdoor, the average company in the US spends about $4,000 to hire a new employee, and it takes about 52 days to fill an open position.
AI Interviews: The Next Big Thing?
It’s understandable, then, that companies would turn to AI as a way of minimizing the damage that this war for talent inflicts on their bottom lines. After all, no organization wants their HR team to waste valuable time on low-level tasks (like 15-minute screening calls) that could be accomplished automatically. Cognitive technology offers these businesses a way of systematizing and scaling a process that, until now, has been pretty labor-intensive.
In the face of rapid growth, that’s an inherently attractive concept. It’s also not a particularly new concept. Companies are already using software to streamline the first steps of the recruitment process, often with mixed results.
For example, Amazon’s experimental in-house hiring tool — which incorporated AI — overwhelmingly preferenced male candidates. It scored resumes that included modifiers like “women’s” less favorably, penalized candidates who had attended all-women’s colleges, and privileged verbs that are more likely to be found on men’s resumes.
As vendors begin testing more sophisticated AI recruitment tools, this sort of bias becomes more of a concern. Talentify, an Orlando company that offers an automated phone-interviewing tool, is testing technology that claims to judge an applicant’s tone of voice and confidence.
Though Talentify’s CEO says the goal is to simply provide more information about a candidate (and not make any final decisions about their fitness for the role), it’s easy to see how AI could potentially discriminate against candidates with certain speech patterns.
That’s troubling, because the consequences of AI discrimination can be devastating on both a personal and social level. There’s also more than a few legal concerns; in the US (and in most countries), the law prohibits discriminatory hiring practices — so businesses are right to be cautious.
The Opinion that Matters Most: The Candidate’s
A more practical concern for many companies, however, is the perception of AI amongst potential hires.
Darlene Racinelli, a financial controller with 30 years of experience, recently sat through an automated phone interview for a Texas manufacturer. Overall, she said, she finds automated interviews frustrating because she can’t ask the interviewer any questions of her own.
Racinelli’s experience reflects a common attitude among modern job-seekers, who see initial interviews as equally important to them as they are to the employer. Screening calls are an opportunity for a candidate to determine whether a position will be a good fit for them. When you can’t ask questions of the screener, it’s harder to find out whether you really want the job.
She also shares another common criticism of AI interviews: they’re, well, boring.
In her most recent automated interview, the system asked her to describe the most difficult challenge she’s faced. This stock question was the last straw.
“At that point, I hit 9 and just ended it,” she says.
Similarly, CEO of Salient Medical Solutions Michael Pizzorno says he’s never had an AI interview — and he wouldn’t ever agree to one.
“I think hiring is a human process, so I think it sends a real questionable message in my mind of how you’re evaluating people,” he comments. “It’s one of those things you can do, but I don’t necessarily think you should do.”
In a tough labor market, opinions like Pizzorno’s matter more than ever. Candidates have unprecedented leverage, insofar as companies have to aggressively compete with one another for talent due to low unemployment. So if potential employees are turned off by having to interview with an AI tool, it could actually work against an organization’s recruitment goals.
Not all job-seekers are averse to the idea of interviewing with AI, however.
Leslie Nienaber, a digital marketing specialist from Cleveland, took an automated phone interview last year. She says that it afforded her more flexibility than a traditional phone screening.
“It was convenient for the time and the fact that I got to prepare a little,” she says. “I knew they could still hear my voice, and hopefully recognize my intent and expertise based on how I spoke.”
For some, the convenience Nienaber describes could be a game-changer. If a candidate works full-time — or perhaps holds down more than one job — scheduling an interview with a human being can be burdensome. Automated tools give them the flexibility to choose times that work with their schedule, even if it’s after-hours or on the weekend. This can be particularly valuable to lower-income workers, who may not have the luxury of taking 30 minutes during the day for a personal call.
How Can You Win In Today’s Tight Labor Market?
Overall, using AI as an interview tool could potentially make the hiring process more flexible and friendly from a logistics perspective. Candidates can schedule interviews at a time that works for them, and HR managers don’t have to sit through hundreds of initial interviews.
But if employers want to use AI to make recruitment more efficient, they’ll have to contend with its shortcomings. From an ethical perspective, there are a lot of potential problems, particularly with bias.
Of course, there’s no guarantee that human interviewers will be fair, either — but there are at least mechanisms for controlling human prejudice in the workplace. Discriminatory recruiters can be reprimanded or even fired. You can’t, however, punish a machine.
It can also be a lot harder to tell when a machine is spitting out biased recommendations. AI is still largely a “black box,” but many simply trust that its outcomes are correct because they are based on data — without considering whether the data itself is flawed (which is often is).
The reality is that recruitment will never be an easy task. Even at the earliest stages, many aspects of the process demand a set of soft skills (like empathy and the ability to make others feel at ease) that AI currently cannot provide. Though some prospective employees may feel okay with that, many don’t. Employers need to actively consider that as they look for ways to hire more intelligently.
That’s not to say that AI is never helpful in today’s competitive labor market. New technologies can make a big impact in another area of hiring: retention. Poor employee retention can be costly; in fact, the total cost of employee turnover is between 100–300% of the replaced employee’s salary, according to the Harvard Business Review.
That’s why the world’s largest retailer, for example, uses OCR technology to ensure they retain employees by continuously assessing what pay and benefits would be most competitive, based on the location and position they’re in. Applications of AI like this can still cut out a lot of manual work and make employers more efficient — but they also aren’t too risky, because they don’t use cognitive technology to make judgments about people.
Intelligent automation is another way to cut costs and make your hiring process easier. Instead of saving time on the front-end — by having AI do interview screenings — you can save your HR department hours of time after the hire, by automating employee onboarding. This not only ensures that you get the right candidate, but also helps create a seamless employee experience. Happy employees equal retention, and retention equals money saved.