Illustration by George Wylesol
The next time you apply for a job, there’s a very good chance that the first person you talk to won’t be a person at all. A growing number of companies are using artificial intelligence (AI) not just to scan resumes and schedule interviews, but to conduct actual job interviews too.While it sounds like something you would expect from a Silicon Valley start-up, a third of respondents to a 2017 Deloitte survey said they already used some form of AI in their hiring process. Urban Outfitters, Under Armour, HBO, FedEx, Goldman Sachs, Walmart, and even the Boston Red Sox have used AI-fueled interviews like those created by Montage and HireVue to recruit, screen, and help job seekers move through the hiring process.
The practice is growing, too. Kurt Heikkinen, President and CEO of Montage, says that currently, “100 of the Fortune 500 are clients of ours.” That’s a lot of potential jobs that require interacting with robots before meeting the humans on staff.It works like this: A job applicant sits in front of their own computer whenever they have free time and video records answers to a series of questions, such as “Explain why your background or experience would be a good fit for this job?” and “Describe one of your greatest professional accomplishments.” Depending on the robot asking the questions, they might also be asked to write an essay, answer multiple choice questions, play little games, or even draw—something Urban Outfitters required at one point.At Unilever, candidates who go through the HireVue process are asked to “solve real-world problems using Unilever scenarios so they can learn more about Unilever and how we do business,” according to Mike Clementi, Vice President, Human Resources, Unilever North America. Candidates have 30 seconds to prepare and three minutes to answer each question or complete each task.The video interview is analyzed by advanced machine learning. An algorithm evaluates facial expressions, word choice, body language, vocal tone, and thousands of data points, which are translated into a score. The robot will then decide whether the person is a good fit for the company. If candidates pass the robot screening, they get to talk to a human, who will ultimately decide whether or not they get the job.
AI isn’t just being used for interview questions. Some companies prefer candidates chat up a chatbot (like the ones devised by Montage, as well as Mya and Karen.ai) as the front line of recruitment, while others, including Unilever, prefer a combination on-demand interviews and the neuroscience-based mind games created by Pymetrics and the like, which claim to evaluate a candidate’s personality for such things as embracing or avoiding risk. At Unilever, the games created by Pymetrics are specifically meant to “gauge cognitive, emotional, and social traits against a Unilever-specific functional profile.”
While companies love AI for vetting entry-level employees, there are problems with using machines. Amazon was recently in the news when it was revealed that an AI hiring tool they were testing was discriminating against women by weeding out candidates that went to women’s colleges among other parameters.Taking the human out of the hiring process is a good thing, Loren Larsen, CTO of HireVue, said. “People often don’t realize how much bias there is in the process and the norm is that people really don’t know what really drives success in a job, or what questions to ask to identify success,” he says. “By using a structured interview process where the questions are carefully designed and the answers are analyzed and compared with hundreds or thousands of samples, we can begin to predict job performance more accurately than human evaluators and begin to remove bias that has always been present in the hiring process.”
Do robots discriminate?
However, as Amazon’s AI lesson showed, algorithms can be biased, too. “If a company's highest performers historically have been identified as white males between 30 and 40 years old—because those individuals were frequently promoted into next-level jobs—that bias can inadvertently become built into algorithms that learn from talent management patterns," the Society for Human Resource Management notes.Unilever’s Clementi agrees, “When using AI in the hiring process, it’s essential to be diligent about auditing and editing the algorithm for fairness. For instance, the algorithm should not be made based on the traits or behaviors of a homogenous group of people. The algorithm needs to be based on traits that are truly predictive of job performance.”While Amazon’s team noticed the issue before implementing the algorithm into the wider company and Unilever, Montage, and HireVue all claim to work to get rid of discrimination in the hiring process—and it may work eventually—as Amazon’s effort shows, it’s hard to root it out entirely, as bias in AI has been well-documented. Currently, there is little recourse for candidates who suspect they’ve been discriminated against by an algorithm.
AI hiring processes can also be difficult for older applicants who may not be particularly tech savvy, as well as for applicants who may not have the financial means to access a private computer where they can sit and do a video interview in peace. Plus, if you sit through an AI interview or play one of the online games, and don’t make it to the next round, an algorithm won’t necessarily give you feedback on your missteps.
That’s where companies like Retorio come in. They hope to fill a niche by helping people game the system—a little anyway—by using their algorithms for good.Marketed as a form of AI personality coaching that can be used for anything from sales training to helping you improve your personal relationships, Retorio lets interview-goers do a practice run before a video interview. It’s software will analyze up to five videos for free based on criteria like speaking energy, sentiment (positive v. negative), eye contact and speaking speed. It even analyzes your personality type. Then it offers AI-based tips on what to work on to ace the real interview.
How to win over a cold-hearted robot
So how do you ace an interview with a robot? Christoph Hohenberger, one of the company’s founders, recommends being expressive, speaking clearly, activating your facial muscles (skip the Botox), and keeping in mind that companies look for keywords, so go ahead and mention the fact that you went to an Ivy League school multiple times during the interview, because AI can’t roll its eyes.Having a blank background for your video interview and making sure you have flattering lighting, getting a decent microphone, and if possible practicing with a friend on Skype or FaceTime are all good ideas too. Act just like you would if you were speaking to someone face-to-face, complete with eye contact, clear enunciation, and appropriate attire.Make sure to do your homework in advance too. Research the company, role, and hiring manage, if possible, and try to assess the critical success criteria. “Candidates should present themselves and market themselves in a way that brings to light their personality, their motivational interests, their understanding of the job, and tie it back to their experience in a way that a resume wouldn’t allow them to do,” says Montage’s Heikkinen.In short, keep it natural as if you spend all day talking to artificial intelligence. Besides, that might be the reality soon.