Bias in the Machine: Will More Tech Help?
Laura I. Gómez, founder of Atipica, a recruiting software company, was sitting on yet another panel discussing the importance of diversity and inclusion efforts, specifically in tech, when something dawned on her. How many more of these panels would she have to participate in before things changed in the industry? She’s worked at Google, YouTube, Twitter and Jawbone, and has become a leading expert on diversity in the workplace, but the rate of change for diverse hiring practices across all industries wasn’t happening fast enough.
Prevalent in tech is a homogenous culture that critics say stifles female and minority opportunity. Looking at the broader picture, tech is hardly alone. To some extent, every industry is struggling to make its recruiting and development practices more inclusive.
Jeanne Achille, chair of the Women in HR Tech Summit at the HR Tech Conference and CEO of The Devon Group, says patterns persist as companies grow and employees enjoy the status quo—a level of comfort that comes with doing “the same thing as everyone else.”
Though it may seem the onus is on HR to “fix” the problem of bias, it’s not that simple, as experts say bias starts taking root in childhood, and is often built into the institutions individuals encounter as they grow—and can be particularly rampant in rapidly developing industries such as tech, where diversity concerns can be lost amid the forward motion. However, there are ways for business and HR leaders to lead the charge—from training to talent management—to break down these barriers.
Beginnings of Bias
Daphne Wotherspoon, managing director of technology contract staffing at HireStrategy, spends a lot of her time counseling clients on the subtleties of language, particularly how it can create bias, starting with job descriptions.
Of course, bias, unconscious or otherwise, doesn’t start in the workplace. It begins much earlier, says Catherine Ashcroft, director of research at the National Center for Women and Information Technology. As we grow and develop, we receive different messages from a variety of sources: the media, our parents, peers and influencers. From these messages, we develop schemas, Ashcroft says.
“Schemas help us determine ‘appropriate’ behavior or jobs for different people,” she says. “When something or someone doesn’t fit that schema, that’s when bias creeps in. Though people think they’re making informed decisions, those decisions are often subtly biased.”
Echoing this sentiment, Phil Alexander, CEO at Nexus A.I., an HCM organization that mines a company’s data to find the best internal candidates for projects, says that humans organize their world by “categorizing.”
“We all have implicit biases that have been shaped by our lived experiences,” says Alexander. “The tendency for most people is to hire, promote or advance someone who feels or appears like themselves.”
One of the many problems with unconscious bias is that it conflicts with someone’s conscious values, he adds. He sees it every day when clients utilize Nexus A.I., which identifies top talent for specific needs within an organization without looking at race, gender or staffing patterns. Nexus A.I. clients have the option of receiving anonymized results, but so far none has requested it.
“Ultimately, by seeing who’s been selected, these employers are adding in bias,” says Alexander. “They still decide who’s going to be on the team, and that human component is biased.”
These implicit biases are a big reason why people of color and women are often underrepresented, particularly in the tech industry. And when bias isn’t addressed, the ramifications can be enormous, both for employees and employers.
Addressing unconscious bias is important for business, both ethically and financially. New research from Catalyst, a global nonprofit organization dedicated to building workplaces that work for women, has revealed that people of color suffer an “emotional tax” in biased, non-inclusive workplaces.
The study, Day-to-Day Experiences of Emotional Tax Among Women and Men of Color in the Workplace, found that nearly 60 percent of 1,600 respondents—multi-racial professionals working in corporate and non-corporate settings—reported being “highly on guard” at work, meaning they were in a constant state of preparedness to deal with potential bias or discrimination. Such employees reported disruptions in sleep, which Dnika J. Travis, lead study author, says can have negative implications for an organization’s productivity.
“If health and well-being of employees aren’t taken care of, companies can’t maximize the talent of their workforce,” Travis says.
Nearly all (90 percent) of the 1,600 respondents who reported being highly on guard were also very ambitious. They want to be influential leaders within an organization, she adds.
“These professional men and women have such high aspirations; if companies don’t leverage it, they’re missing out on critical talent,” Travis says. “Leaders need to hold themselves and their organizations accountable to mitigate emotional tax and make a difference in creating inclusive environments.”
Research from the McKinsey Global Institute found that companies ranked in the top quartile for racial and ethnic diversity are 35 percent more likely to have financial returns above their respective national industry medians. Those in the top quartile for gender diversity are 15 percent more likely to have better financial returns than the respective national industry medians.
Though it shouldn’t be the driving focus, the business case is there for tackling bias and ramping up inclusion. But workplaces appear to be dragging their feet, and one of the most problematic industries is tech.
Troubles in Tech
Data from the NCWIT show that 26 percent of U.S. professional computing occupations are held by women, which is in stark contrast to the proportion of overall professional occupations held by women (57 percent).
“The creators [of tech] are few, and the consumption is large,” says Gómez. “We need to think about who’s creating the technology and who’s consuming it. When it stays in the hands of a homogenous group, that group reaps the benefits while everyone else is left behind.”
This creates an ethical problem, but also a staffing issue. ISACA, a global nonprofit that develops and advocates for best practices in IT governance, reports that globally there will be 2 million unfilled cybersecurity-professional jobs by 2019.
While some experts say there’s a talent shortage within the industry, others counter that unconscious bias is impeding smart hiring decisions.
Ashcroft says bias can be overt and conscious, but most often, it’s subtler.
“We, as humans, construct a world that fits our needs,” she says. “If a homogenous group is building a team or starting a company, then those needs are limited to that group and that group only.”
Pooja Jain-Link, vice president and associate director of research at the Center for Talent Innovation, says addressing this group-think problem is one of the main challenges in tech.
“There is such a strong male presence in tech that biases are amplified,” says Jain-Link. “One of the biggest problems with hiring people who are like you [the dominant group] is that you’re blinded to other needs in an organization. Research has shown that if you don’t personally think there’s a need for something, be it diversity or inclusion, then you will never understand why it really is necessary.”
Another issue in tech is the speed at which the industry is growing and changing. According to Imo Udom, CEO of WEPOW, a video-interviewing platform, tech companies often view success as how quickly a company can grow and/or meet goals. This velocity puts diversity, inclusion and other issues on the back burner. Udom says the mentality of “growing at any cost becomes a justification for dealing with other problems later.”
As evidenced by numerous studies, “later” still hasn’t come for many businesses. With that said, some large organizations are leading the charge in diversity and inclusion practices.
Lead By Example
Nancy Testa, chief diversity officer at American Express, takes diversity and inclusion seriously, as evidenced by the amount of company-wide training and programs dedicated to it. In 2015, American Express partnered with the NeuroLeadership Institute and created the DECIDE program. The initiative was designed to help employees recognize unconscious bias and understand how it impacts decisions in hiring, recruiting and promoting talent.
Testa says the program had a tremendous impact: 98 percent of participants said it helped them understand unconscious bias and how it influences daily decisions.
This year, the company is requiring every vice president and C-level executive, up to the CEO, to complete an inclusive-leadership experience, which focuses heavily on unconscious bias, Testa says.
“It’s important for leaders to be educated about this because they role model our company values,” she says. “When they take this knowledge and mentor other leaders, that’s when we’ll continue to see the change we’re looking for.”
Testa also emphasizes the importance of employee networks. American Express has 16 networks with more than 100 chapters globally, including the Black Employee Network, Disability Awareness Employee Network and Women in Technologies Employee Network. According to the company, employee networks broaden participants’ community of colleagues, as well as their knowledge of the business by driving employee engagement and promoting a culture of inclusion.
Ken Bouyer, the Americas director of inclusiveness recruiting at Ernst & Young, says EY has received praise for its diversity and inclusion efforts, including its People Leadership Matters training, which encourages employees to pause before making key decisions.
“We teach attendees to … ask themselves, ‘Is it a preference, tradition or requirement for a role or opportunity?’ This question allows us to reflect and remove the potential for unconscious bias,” says Bouyer, who says such training has been personally impactful. “Before undergoing unconscious-bias training at EY, I thought I was Mr. Fairness—I treated everyone equally. The trainings we’ve implemented have opened my eyes to my own biases.”
EY also provides recruitment training, which helps recruiters define bias and explains how it can creep into the recruiting process. Additionally, newly promoted leaders undergo training at the yearly Milestone Event, which has a two-fold purpose: to celebrate the promotions and develop these new leaders. Each year, EY invites Dr. Mahzarin Banaji, a leading expert in unconscious-bias studies, to speak at the event.
Banaji explains to the leaders that everyone has bias—and accepting it is part of the process of addressing it.
“Unconscious bias is a scary topic for folks,” says Bouyer. “We [HR leaders] know it exists, but if you want to move your organization to the next level, you need to have the courage to hit this problem head-on—you have to be comfortable with being uncomfortable in this space.”
Addressing the Human Component
Training and education are important in addressing bias, but without the right tools and consistent follow-up, those efforts often fall flat, say the experts.
The first crucial tool, according to Catalyst’s Travis, is simply to talk about bias—listen to employees who are at risk for emotional tax and break through possible “roadblocks” to the discussions.
Travis says people want to feel valued, heard and seen in an organization, and these three factors are critical in creating an inclusive environment. On the heels of listening, she adds, is learning, which includes conducting climate surveys and continuously collecting data on the organization’s demographics and comparing it to national benchmarks.
“HR leaders need to embed a focus on unconscious bias and intersectionality [of race, gender, ethnicity] into the fabric of the talent-management process—they should be able to identify gaps in intersectionality, from recruiting to onboarding,” she says. “Leaders need to ask themselves, ‘What are we doing to ensure that everyone is valued, heard and contributing to the organization?’ and, ‘How are we ensuring people stay?’ ”
At Atipica, Gómez’s company, data mining and analysis play a key role in identifying areas for improvement within an organization’s hiring pipeline. Atipica is built on top of a company’s existing talent database and predicts hiring needs while unveiling recruiting and hiring patterns. These patterns ultimately help bring bias to light.
“Unpacking bias is complicated and multi-layered, particularly because humans have inherent historical behaviors that have been accepted as the norm—these norms are harmful because they’re biases,” says Gómez.
Something Udom at WEPOW has kept in mind over the years is a quote from Mitch Kapor, CEO of Kapor Capital: “Genius is evenly distributed across all zip codes, but opportunity is not.”
“In tech,” Udom says, “we preach meritocracy; we want the best people for our teams. That means we’ll look for candidates from Stanford or Berkeley, but if someone from a disadvantaged group applies who didn’t go to a prestigious school, they’re already considered ‘not as great.’ That’s bias—not every smart person goes to an Ivy League school.”
WEPOW helps companies address potential bias in the recruiting and interview process by allowing for structured, personalized video interviews. Most experts agree that unstructured interviews are rife with potential for bias.
“Pre-recorded interviews mean the questions and interview experience are identical for all applicants,” Udom says.
Once a candidate records a response video, multiple people on the hiring team review it. “This allows for greater transparency among recruiters, who may have previously relied on a ‘gut feeling’ to turn away certain applicants,” he adds. “Now, they need to justify their decision.”
Udom also stresses that not all candidates have strong resumes and, by relying solely on applicant-tracking systems to pull out key words, organizations are missing out on talent.
“[Members of] under-represented groups often don’t have great resumes, or they have untraditional work histories. Conducting video interviews allows applicants to share their story and explain how they can add value to the organization—they’re given an opportunity to show that they’re so much more than words on paper,” he says.
Tigran Sloyan, CEO of CodeFights, a skills-based recruiting platform, agrees.
“Only 2 percent of engineers have an opportunity to work at places like Google or Facebook,” he says. “But a lot more than 2 percent of people have the skills and abilities to be great.”
CodeFights Recruiter was designed to help employers find the right talent based on skills assessments, not pedigree. It matches employers with technically pre-screened candidates and provides advanced assessment tools for testing and interviewing candidates.
“We need more structure around the interview process because skill is an objective measure,” Sloyan says. “By layering artificial-intelligence data with human data, we begin to see inconsistencies and can determine where bias exists. Only when this happens can we effectively source, hire and promote candidates based on skill, not biased, human-driven feedback.”
Achille of The Devon Group cautions against relying solely on vendors and AI to “solve” unconscious bias.
“Unconscious bias starts with diversity and inclusion efforts,” she says. “Begin at the board level and ensure that they not only understand why diversity and inclusion are important, but that they fully support it. These ideas need to be baked into the leadership mantra and overall culture of an organization.”