Have you ever lied during a job interview? Most of us have, at least a little. But next time around, artificial intelligence may be watching your face’s every move, assessing the honesty of your answers, as well as your emotions in general. It may also try to determine whether your personality is a good fit for the job.
That’s already been happening for more than a year to applicants for entry-level jobs at multinational corporation Unilever, maker of Dove soap and Lipton tea among many other consumer products. Applicants, who often find the company’s job opportunities through Facebook or LinkedIn, can skip uploading their resumés and simply use their LinkedIn profiles if they wish. They then spend about 20 minutes playing a dozen neuroscience-based games intended to evaluate their personalities for such things as embracing or avoiding risk, to see if their personalities are a good fit for the particular job.
Then they perform a video interview, with preset questions, which they can do on a smartphone or tablet as well as a computer. That’s where AI comes in, measuring their facial expressions to capture their moods and further assess their personality traits. (Does the AI also evaluate their truthfulness? The company isn’t saying, but given the vast body of knowledge about spotting truth or lies from facial expressions, you have to assume that it is, or at least that it’s evaluating honesty as one of the candidate’s personality traits.) The software, provided by a company called HireVue, selects what it believes are the best matches and returns those to a human recruiter, along with notes from the AI about what it observed in each candidate. Many other major employers, from Dunkin Donuts to IBM to Carnival Cruise Lines to the Boston Red Sox are using the service as well.
A good thing?
Believe it or not, there are a lot of advantages to this approach, both for the employer and the candidate. For one thing, it creates convenience on both sides, allowing candidates to be interviewed whenever they want, and recruiters to review those interviews on their own schedules. Unilever also claims this new approach has contributed to ethnic diversity, with what it calls a “significant” increase in non-white hires. (The company says it was already at parity when it comes to gender.) This makes sense: Human recruiters, however well intentioned, are subject to unconscious bias that properly trained AI may be able to avoid.
Perhaps most importantly, the new approach increased the number of colleges Unilever hired from for entry-level positions from 840 to 2,600. That’s significant, because major employers with a finite number of college recruiters tend to send them to the most elite colleges. That means they’re mainly recruiting among candidates whose parents can afford those schools. So the new method of hiring leads to greater socioeconomic diversity as well.
Most of all, it saves a lot of time for both Unilever and the applicants, which may mean the company gets to grab great job candidates before the competition hires them.
On the other hand…
There are some big potential downsides to this approach as well. For one thing, it’s a little creepy. Are job candidates told that their facial expressions will be analyzed by algorithm? It doesn’t seem to me there’s any legal obligation to tell them, and some of them may assume it will happen or may have read press accounts of the practice. Even so, the idea that AI is analyzing your face to determine your fitness for a job is a little distressing, even if that sort of analysis is becoming more and more common in our modern world. As one facial-recognition expert told the Financial Times, people who know their facial expressions are being analyzed tend to display self-consciousness. (Wouldn’t you?)
Second, our moods are a very changeable thing, and some of those changes may or may not have anything to do with a job we’re applying for. Someone whose relationship recently ended or who has, say, a toothache may display signs of unhappiness, anger, or confusion that have nothing to do with that person’s usual personality or fitness for a job. In a human job interview, you can choose to tell the interviewer if there’s something going on that may affect how you come across. You can say it in a video interview as well, but the AI might screen you out before a human ever gets to hear it.
Then, there are a wide range of physical issues that could potentially affect our faces and facial expressions. What if the person being interviewed has facial tattoos? Facial scarring? Suffered a stroke? Had Botox injections? Again, a human interviewer might see those things and take them into account, but would the AI screen such candidates out so that a human hiring manager never watches the video?
For employers, especially those trying to hire a lot of people quickly, this could be a very useful tool, and perhaps a great way to boost diversity as well. But for job candidates, knowing your emotions will be read, it’s a good reason not to apply for any job or to any company you’re not genuinely enthusiastic about. Or it may be a good reason to brush up on your acting skills.
Here’s a look at AI analyzing emotions: