Here’s a fact: when you’re job hunting, human beings aren’t always looking at your résumé.
Experts used to say recruiters spend six seconds reviewing your résumé before tossing it in the trash (grab their attention fast!).
But forget that idea—computers have been doing the gatekeeping for years, screening for keywords and employment gaps. Now tech is taking things to the next level.
Much the way Netflix suggests Unbreakable Kimmy Schmidt after you watch one 30 Rock episode, companies are starting to use artificial intelligence (AI) and sophisticated tests, in addition to algorithms, to not only find candidates for a job but also predict who will kick ass at it. The hope is that these tools will reduce costly turnover and, most important, help fix the human error and bias that have historically shut out women and minorities.
Great goal. But how will this all work?
After gathering company data, “these systems would be trained in ways that reflect ideas of ‘successful’ and ‘unsuccessful’ at a given company,” says Meredith Whittaker, cofounder of AI Now, a research institute at New York University that studies the social effects of AI.
“But if a company isn’t diverse, especially in executive and leadership roles, [the software] would likely replicate existing biases that see certain people, often white men, as inherently more capable and successful than others.” In other words, version 2.
0 of the hiring bots may be an improvement—or maybe not. But they’re coming.
At least 98 percent of Fortune 500 companies use an automated applicant tracking system (ATS), or résumé screener, according to Jobscan, a tech company that helps applicants improve their résumé to beat the machines. “ATS software is designed to assess an applicant based on keywords,” says Amanda Augustine, a career advice expert at TopResume, a New York City résumé writing company.
Without the right ones, she warns, your résumé “may never get past this initial digital gatekeeper.”
A simple hack: “Pick out the important keywords located in the job listing and incorporate them in your ‘Skills’ and ‘Work Experience’ sections,” Augustine says.
Not sure which are the most important ones? You can make a word cloud out of a job description using a free tool like Wordle, she suggests.
“A key issue is gaps in employment,” says Ifeoma Ajunwa, Ph.D.
, an assistant professor of organizational behavior at Cornell University. Some algorithms, Ajunwa explains, “might be trained to reject résumés with gaps, which could mean that women who take time off to raise children may want to explicitly list that or some other part-time activity on their résumé rather than leave those years blank.
” Augustine says to consider proper formatting too, because complicated design elements can get lost on the software. “Stick to a simple design that doesn’t include embedded charts, images, or unusual fonts.
Make the predictive algorithm predict you.
Companies are also moving past traditional hiring assessments like personality tests (which experts say don’t accurately predict your job success) and are instead using AI, algorithms, and neuroscience games to better match your abilities to a role.
Frida Polli, Ph.D.
, created Pymetrics, an AI-based recruiting platform, to fix the hiring process after seeing her fellow Harvard Business School classmates struggle to find the right jobs. “I was amazed to see how many people put a lot of time and effort into getting recruited into a new field and then were not happy when they actually started working in it.
” If they’d gotten a better understanding of their fit for that role, she says, some of that pain—for workers and companies—may have been prevented. Polli says Pymetrics uses algorithms to predict whether someone’s skills line up with a particular job by comparing their answers with those given by a company’s top performers.
Platforms like Pymetrics also say their software can also decrease hiring biases, such as discriminating against résumés with female or nonwhite-sounding names, by focusing on selected traits that social science research says are largely free of gender and racial bias, like risk tolerance and planning style. “If a hiring manager puts the most emphasis on hiring from a résumé, women and people of color have a reduced chance of getting the job,” Polli says.
The method appears to be working, at least for some companies. While Pymetrics doesn’t disclose client names for privacy reasons, it has reported impressive gains in 2018 hiring: an 18 percent increase in female technical hires and a 20 percent increase in minority interns at one financial service company.
But some industry professionals are wary. “A lot of algorithmic hiring tools use data from an employer’s existing network or what’s available on the Internet,” says Stephanie Lampkin, CEO and founder of Blendoor, whose hiring technology aims to reduce unconscious bias by hiding data like names and ages.
“If you rely solely on limited information, you’re not accessing underrepresented talent.”