Implementing Video Interviews? Here’s How To Beware Of Bias
Just two weeks before the U.S.’ first known case of COVID-19, the Wall Street Journal’s Hilke Schellmann forecast that hiring processes powered by artificial intelligence were just around the corner. A desire to stay competitive in an increasingly tech-reliant world was the force that she predicted would usher companies into a world of recruiter chat bots, AI-powered background checks, even machine learning services that study the language in a candidate’s application materials in order to determine whether they have certain personality traits.
Now, there’s a new, more urgent force propelling that evolution: In the short term, companies that are still hiring through the novel coronavirus need to move the hiring process forward without the ability to interview or train new employees in person. Hiring managers may rely on phone interviews, but especially with the rapid adoption of videoconferencing for every other form of human communication, many will turn to video technology.
But as with face-to-face interviews, those hiring need to make sure to avoid unconcscious basis that could lead to discrimination — not just on their part, but on the part of the software system being utilised too.
“Utilizing video interviews over the past five years, our team has always seen this as a differentiator,” RealREPP CEO and founder Johnny Renaudo said. “In some cases — and I think we’ll be seeing this more — it’s just more efficient.”
In hiring terms, “video interviews” can mean several different things. First, there are the standard videoconferencing tools such as Zoom, GoToMeeting and Skype via which a hiring party may choose to have an interview with a candidate as a more personal alternative to the phone.
Some companies turn to software built just for videos in the hiring process, offering more advanced features like integration with applicant tracking systems, automatic scheduling and other features, and there are dozens of platforms to choose from.
Renaudo said real-time videoconference interviews shouldn’t be a dramatic change for hiring teams: "I don't think video interviewing is so different from in-person interviewing, to be honest with you," he said. "Video quality is so good, so clear and so personal. The format drowns out the background and it focuses on just you two."
But another feature offered by many video interview platforms is asynchronous video interviewing. In this case, the candidate records the answers to a predetermined set of questions on their own schedule for the hiring party to evaluate later. This allows companies to grow their teams at great speed and scale, and it circumvents planning and scheduling, making it easier for the hiring process to move forward despite busy or conflicting schedules.
When CVS announced on March 23 its plans to launch “the most ambitious hiring drive in the company's history” and immediately hire 50,000 full-time, part-time and temporary employees to staff stores across the U.S. in response to the coronavirus, the “technology-enabled” hiring process it described at the scale and speed needed may well involve asynchronous video interviews — perhaps even screened initially by AI smart enough to make the first cut.
Beyond speed and scale, according to on-demand video interviewing platform HarQen President and CEO Tim Ihlefeld, asynchronous video interviews have another big advantage: They may help hiring teams reduce or eliminate potential interview bias by providing structured, consistent interviews across all candidates for any given position and ensuring interviews are conducted identically, which can’t be the case with a live human interviewer.
For example: “There is no opportunity in an automated, pre-configured interview to ask candidates differing questions that may stem from personal, unconscious biases,” Ihlefeld told Recruitment Innovation Exchange.
But artificial intelligence is not immune from bias and in fact, many hiring tools have come under fire for discriminatory results, to the extent that in 2019, legislators including Sen. Cory Booker (D-NJ) introduced a bill, the Algorithmic Accountability Act, which would authorize the Federal Trade Commission to create regulations to check AI decision-making tools, like those used in recruitment and hiring, for bias.
Leading up to the bill, privacy rights nonprofit Electronic Privacy Information Center filed a complaint of potential bias, inaccuracy and lack of transparency with the FTC against AI-powered video interview platform HireVue. Also in 2019, the Department of Housing and Urban Development charged Facebook with violating the Fair Housing Act by allowing advertisers to discriminate based on race, religion and disability status. And in 2018, Amazon was pressured to scrap an AI-powered recruiting tool that displayed a bias against women.
Ihlefeld’s recommendations to combat bias include vetting and reviewing video interview questions in advance, by committee if possible; auditing and reviewing recorded interviews as they arrive or in a timely manner; and taking care not to use platforms with facial or voice inflection analysis tools that have been questioned or that haven’t been vetted.
Back to the seemingly simplest option, conducting interviews over Zoom, GoToMeeting or other standard videoconferencing software may introduce bias as well, if a company is rolling out new interview processes in response to the coronavirus without ensuring consistent use by all recruiters or hiring parties.
Behavioral economist and Harvard University professor of business and government Iris Bohnet wrote for the Harvard Business Review that structure and consistency in interviews are critical to helping hiring parties avoid or reduce bias. Roll out new interview tools and processes consistently, and once they are in use, Bohnet recommended asking candidates each the same questions in the same ways and evaluating them by comparing their answers horizontally for greater consistency.
Ihlefeld added that no matter what type of video interview is in use, the interview process should always have a human touch.