Associated Press
Artificial intelligence might be biased, but maybe it's less biased than we are.
That's one possibility that doesn't get as much attention in the conversation about how AI could retool our workplaces, especially when it comes to hiring.
It's possible that AI, working as a gatekeeper, will do a better job than humans at reducing bias in how we get jobs. That could make it more likely that skills — and not fancy degrees or connections — play a bigger role in who vaults ahead in the job market in the coming years.
Finding ways to reduce barriers to who gets a fair shot at certain jobs is the hope of Andreas Leibbrandt, an economist at Australia's Monash University. He told Business Insider his research indicates AI can be more evenhanded in evaluating job candidates in a male-dominated field like tech, where men lean toward hiring other men.
That doesn't mean humans shouldn't make the final call about who gets the job offer, Leibbrandt said. He sees AI, as a tool, like a knife: "If you don't know how to use it, you're going to cut yourself, but if you learn how to use it, you'll be able to make food quicker," he said. "That's how I view it. I think it can radically transform the recruitment space."
Women seeking tech jobs seem to be on board with AI as a surrogate recruiter, according to Leibbrandt's research. "The one thing that really pops out is a strong sentiment amongst women that they are confronted with bias when assessed by humans," Leibbrandt said. "Their concern is much less so when they are being assessed by AI."
Leibbrandt looked at the tech sector because anecdotal evidence from many women showed they experienced bias when going for jobs. And there are, he said, perceptions of "bro culture" in that environment. The research indicates women might be onto something.
"We do observe that when recruiters know the gender of the applicants, then they discriminate against women. When we remove the gender, there is no gender difference," Leibbrandt said.
To test how AI would handle job seekers, Leibbrandt and his colleagues placed ads for actual tech jobs and included a line that indicated that either a person's application would be reviewed by a recruiting team or by AI. When AI was in the driver's seat — at least with the initial screening of an application — women were more likely to put in an application. The implication, Leibbrandt said, was that women saw completing the application as worth their time.
"Women are less likely to drop out when provided with information that they are being assessed by AI as compared to humans," he said. Leibbrandt said it doesn't mean women don't worry about bias with AI. They just think, broadly speaking, that it's less of a concern than bias from people, he said.
For the men in the experiment, Leibbrandt noted, the opposite happened when they were told their application would be scored by AI and not by a recruiting team: They were less likely to complete an application.
"It almost seems like that a privilege has been removed when you have AI instead of human recruiters," he said, referring to male applicants.
There are other ways AI could intervene in the process. The tech could also help by scoring candidates' qualifications. When that happened in the study, male hiring managers didn't discriminate against women.
Leibbrandt said other research indicates that a similar pattern plays out when it comes to race. Those studies are somewhat constrained by small sample size, he noted, but early indications are that non-white women, for example, might feel they have a fairer shot at a job when AI does the first pass on a pile of applications.
Others are working on ways AI could improve hiring. Barb Hyman is the founder and CEO of Sapia.ai, which uses AI-powered text chats to screen job applicants. It's a departure from tools like the ones where an applicant records a video answering various questions or shows up for an in-person interview.
Hyman, who's worked as a chief human resources officer, said an AI chat that's not timed puts far less pressure on people and can make it easier for people who might feel intimidated by a video interview or as though they're going to be passed over because of their background.
"The differences between a blind application process versus something where I have to do a video or I need to show up in person is very significant for a large group of candidates," she told BI.
Hyman said having that initial screening done by AI can make it so that hiring managers give those candidates who make it through the process a fairer review.
"I can't train you out of your unconscious biases because they're unconscious. So if the market generally wants to embed fairness and achieve diversity, you have to use AI," she said.
Drawing in more applicants can also help employers that are still having trouble finding workers given the low employment rate, Hyman said.
"There's not enough talent so you have to find a more effective way to see talent, to see the potential in everyone," she said.