Can AI Make Hiring Fairer And More Efficient?
14 August 2019
rtificial intelligence applications like Siri and Alexa wake many of us up. They tell us the weather, give us directions, answer our questions, sing us songs, even tell us jokes. They have been welcomed warmly into homes. When it comes to the workplace, people tend to be less welcoming.
A lot of Americans — about 60%, according to one survey— worry about robots taking over jobs. It’s not an unfounded fear. According to Brookings Institution, about 36 million Americans (roughly one third of workers) work in jobs where 70% of the tasks can be automated. AI has found its way onto factory floors and into offices, and the HR department is no exception.
Not only is technology changing the way we do our jobs, but it’s also changing how we apply for those jobs and whether we are considered.
According to a survey released last year by consulting firm Korn Ferry, only 25% of recruiters had yet to use AI in their job. A majority of the respondents (87%) said they were excited about using AI more in the future.
Among the usual exhibitors at the annual conference of the Society of Human Resource Management (SHRM) in Las Vegas in June — like payroll companies and benefits providers — were numerous AI companies. Booth after booth, attendees browsed software that promised to help sort resumes, conduct interviews, track facial expressions, simulate work scenarios, keep track of internal complaints and more.
This does not mean that there will be a bunch of robots sitting around the office going through interviews. Most of this AI is software.
“I think we all jump to this idea of 2030 or 2040, robots walking around, but I use AI to do my job more effectively,” said Amy Cappellanti-Wolf, chief human resource officer at Symantec and a speaker at the conference.
She said she uses AI to gather and analyze data and to remove bias from recruiting. According to Cappellanti-Wolf, the use of AI in HR is now essential, particularly as departments are tasked with doing more with less.
“It’s no longer kind of nice to have, it’s really table stakes for how you are going to be able to run an effective HR and business organization,” she said. “We often come under pressures around being more productive and managing our budgets, and that doesn’t always mean more people.”
Symantec also uses AI in its quarterly survey of employees. “It allows us to really rapidly review where people are at … in their desire to stay or go,” said Capellanti-Wolf. “We can actually predict when people are going to leave, six months before they do it.”
The industry as a whole seems to be adapting and adopting.
“It’s happening. I’m practical,” said Johnny C. Taylor, Jr., SHRM president and CEO, adding that two years ago, there were no conference sessions on AI. That was not the case this year.
A presentation by an India-based company that specializes in applying AI to HR, Aspiring Minds, for example, detailed the use of AI to score video job interviews based on body language, voice emotions and facial expression.
But that kind of technology can make job seekers uneasy. A survey of more than 2,000 adults conducted earlier this year revealed that majority weren’t in favor of the use of AI in hiring. Just 31% said they felt it was acceptable for HR to use AI.
However, it’s not only HR executives who stand to benefit from the use of AI, according to Aram Lulla, general manager at executive recruiting firm Lucas Group.
“The flip side is that you can get noticed and identified because of AI,” Lulla explained. “Let’s say 1,000 people apply to one role. You may only have a subset of one or two people reviewing those resumes, and you might never get looked at. But because AI is there, it can handle so much more volume. You might actually get found now.”
Experts interviewed by Marketplace also argued that using AI to sort through candidates’ resumes helps eliminate bias and ensure that the candidate pool accurately reflects the demographic make-up of the labor force.
Chances are that many candidates have unwittingly used AI in their job search, by using the professional networking site LinkedIn.
LinkedIn collects data from its 645 million-plus members’ resumes, job postings and searches by both recruiters and candidates.
According to John Jersin, LinkedIn’s vice president of product for talent solutions and careers, one of the reasons job candidates frequently don’t hear back is because they apply to jobs they are not qualified for. That’s one thing that LinkedIn wants to help candidates avoid.
“We’ve developed an increased ability to predict which jobs you’re actually likely to get,” he explained. “We’re showing you those jobs preferentially and so you should be applying to more of the right jobs and therefore more likely to get them.”
Applicants on LinkedIn are now informed whether they’d be a top candidate, as well as how many other people have already applied for the job in question.
That’s not to say that LinkedIn’s foray into AI has been without hiccups. The company originally showed recruiters the best candidates for their jobs, even if the candidates were overqualified.
“You would think that recruiters are pretty smart about who they’re going to be able to hire, who’s the right person to reach out to but, in fact, what was happening was most recruiters were reaching a little bit too much,” said Jersin. “They were reaching out to the same set of very, very qualified people. So those people felt overwhelmed with the number of opportunities, to the point of frustration.”
LinkedIn eventually stopped showing recruiters the best candidates and started showing them “who we thought they would actually hire.”
Jersin says that while AI in recruitment is still far from perfect, it’s here to stay.
“AI is a long-term trend,” he said. “We’ve been talking about it for years and we’re going to continue talking about it and working on it for years.”