Skip
Current Issue
This Month's Print Issue

Follow Fast Company

We’ll come to you.

4 minute read

Leadership

This Interviewing Platform Changes Your Voice To Eliminate Unconscious Bias

Hiring is fraught with unconscious bias, but Interviewing.io aims to change the game at the outset to allow only the skills to shine.

This Interviewing Platform Changes Your Voice To Eliminate Unconscious Bias
[Photo: Flickr user Jon Callow]

Unconscious bias is often hiding in plain sight during the hiring process. We know that simply by being human, recruiters and managers can fall prey to signals that suggest the candidate is somehow not part of their tribe. Everything from an ethnic-sounding name to checking a gender box on an application can disqualify a candidate at the resume stage, according to several studies.

Apps such as Blendoor hide a candidate’s name, age, employment history, criminal background, and even their photo, so employers can focus on qualifications. Interviewing.io—currently in private beta—is an interviewing platform that is taking the concept of anonymizing one step further. Even the voices are masked.

This tackles another important unconscious bias at play in the hiring process. One study revealed that a female job applicant’s vocal tone leads recruiters to believe they are less competent. Other research found that "vocal fry," a low-pitched or "creaky sounding" speech common among young American women, is perceived as a negative and hurts applicants’ chances in the job market.

Aline Lerner is one of the cofounders of Interviewing.io. As a woman in tech, Lerner says she’s been one of the lucky ones who hasn’t felt particularly discriminated against. "The closest I came was when I went to MIT and people said to me, do you think you got in because you are a girl?’" Lerner says it wasn’t great to hear, because she would never know if they were right. "I would have preferred to get the things I got based on merit, not gender," she says.

In her years as a recruiter, she was responsible for vetting candidates for the likes of Airbnb, Dropbox, Asana, Lyft, and Udacity, among others. Time after time, Lerner says, applicants were taken out of the running before they even got the opportunity to showcase their coding chops because their resume wasn’t filled with experience at other major tech companies or their degree was from a less-than-elite university.

Interviewing.io started as a platform that would allow tech professionals to take on a coding challenge with an interviewer, as if they were working on a virtual whiteboard. The interviewee’s name and other identifying information was taken out. At first, Lerner says, everyone was given superhero names, but that quickly became problematic. "Not only would we have run into copyright issues," she explains, "But most end in 'man', and that’s not what we are trying to do."

Gender neutral animal names are now given to both the interviewer and the interviewee. And while the two can exchange comments via the interactive text feature on the platform, there is also an option to chat using a voice feature. Here’s where it gets interesting.

Lerner explains that Interviewing.io uses both Twilio cloud communications technology for the call and their own proprietary voice software. The latter has the capability to make men sound like women or vice versa, or to make either gender sound androgynous. "We let companies choose what they want," she notes, adding "Generally we haven’t found any trend as to which is better." Lerner says that when modulating men’s voices, "We add synthetic elements, so you don’t know if it’s a woman, or if it is a man whose voice is processed."

Even so, voices aren’t completely unrecognizable. "There is nothing to take away an individual affect," she says, so the transformed voice still sounds like you.

We did a little mock interview to see exactly how it worked. In the first segment, you can hear us exactly as we sound: two women.

In the second, our voices were modulated to make us sound more masculine. Our identities were masked as we were given handles. Lerner is "Defenestrated Spider" and I was "Grey Narwhal."

Despite the anonymizing, our natural inflections save us from sounding like we have robotic voices. This handily eliminates the possibility that either the interviewer or interviewee would feel like they were talking to a bot.

Lerner underscores the fact that she isn’t advocating for completely blind hiring. Indeed, the candidate and the interviewer can reveal their identities during the course of the interview, should they wish to continue the process. Instead, Lerner says, "The way we filter candidates at the very top of the funnel is broken." Interviewing.io provides a platform that equalizes that very first part of the process. "We want people who are good to show what they can do," she says, without having to discounted for not having a coveted tech pedigree.

Finding culture fits among new hires is just as important, Lerner maintains. "You want someone who is going to be happy, share your vision and gets along with you while spending 8, 10 or 12 hours a day," she explains. But, says Lerner, "Filtering people [by merit] before complicated issues of fit is much more valuable than where they went to school or worked before, or what their name is," she argues.

With the site still in private beta usage among roughly a dozen companies, Yelp has been the pioneer in using Interviewing.io’s voice feature, in addition to its interactive coding space. Lerner says 19 candidates have interviewed so far and of those five or six have received offers, and one has accepted.

Lerner declines to give additional placement numbers because, "this is a very crowded space," and placements equal money made by the company, and she’s not ready to reveal that.

What she does tell us is that about one out of three candidates overall using the platform make it past this initial screening and on to in-person interviews. Says Lerner: "This is much more efficient." Not to mention, much more of a meritocratic process.

loading