Inside the offices of ThirdLove, an Internet retailer in San Francisco, a team of data scientists build algorithms using artificial intelligence to sort and dissect data and patterns among online shoppers. Yet the startup faces an unusual diversity problem for Silicon Valley: It can’t seem to hire enough men.
Nearly 85% of ThirdLove’s 355 employees are female, and women make up nine out of 10 members on its data science team. The startup’s head of data and its co-CEO are also female.
Granted, ThirdLove’s female-oriented workforce makes good sense. The company uses AI, data, and algorithms to help shoppers determine the best fit for a bra. And most men aren’t interested in working for an undergarment retailer even if it has a high-tech bent — or at least they’re not passionate about better fitting bras, ThirdLove’s criteria in hiring. Says co-CEO Heidi Zak: “We have to look hard for men in order to bring in diverse backgrounds and skillsets. But having more women attracts more women.”
Still, ThirdLove is an anomaly in the technology industry, where women’s employment has steadily declined since 1991, after it peaked at 36%, according to the National Center for Women and Information Technology. And when it comes to data science and artificial intelligence employment, ThirdLove is even more unusual. Men make up 88% of all AI workers and 80% of all AI professors.
“Right now, there’s a major crisis in diversity in AI— worse than the rest of tech,” says Tess Posner, director of AI4All, an organization working to get more women and minorities into the industry. “As AI becomes more ubiquitous, taking on a lot more decision making, it’s being shaped and programmed by a single homogenous group.”
The AI field, which is overwhelmingly white and male, is at risk of replicating or perpetuating historical biases and power imbalances through technology. We’ve already seen cases of AI bias, including chatbots adopting hate speech and Amazon’s tech failing to recognize people with dark skin.
Diverse views in AI and in data science is vital going forward, says Sarah Aerni, Salesforce’s director of data science who has championed diversity in data science. It’s a scientific job, but it also involves interpreting information through your own personal lens. “The way I interpret and investigate data and models is different from another data scientist. If I want fuller exploration and understanding, I need diverse perspectives.”
It’s not just bias in algorithms at stake, diversity in tech lends to a more comfortable work environment for women, says data strategist Lillian Pierson, who has spoken out about the gender gap in her field. In her own career, she has witnessed unequal pay, sexual harassment, and problems of women not being taken seriously in a technical field— especially overseas. “You have to put your dukes up to advance and get a raise,” she says.
Having more women in leadership and on corporate boards helps, and we’re making slow but steady progress in that arena, says Dawn Belt, a partner at law firm Fenwick & West and coauthor of the annual Fenwick & West Gender Diversity Survey, which reviews leadership and board diversity in Silicon Valley. “Seeing women in leadership makes women more inclined to take a job.”
Last year—dubbed by some as “the year of the woman” overall— did, in fact, show improvements in terms of gender diversity— at least in board and leadership roles in Silicon Valley. California became the first state to require public companies to include women on their corporate boards, and indeed a record number of them were appointed to boards on S&P 500 companies. Today, women hold 25.5% of board seats at Fortune 500 companies compared to just 15.7% 15 years ago.
Zak, who started ThirdLove with her husband, David Spector, says she hopes that ThirdLove, though still small, will become a breeding ground for women in technology. The firm, which has raised $68 million from investors, has a strong culture of mentorship that encourages helping each other and encouraging leadership opportunities for employees.
ThirdLove data science director Megan Cartwright said that prior to joining ThirdLove she was often the lone woman on all-male data science teams and often felt that she was missing out on the relationships built during after-hour drinks that she couldn’t attend because she was a working mother. A good chunk of ThirdLove’s workforce are parents, so instead of drinks after work, they meet for lunch or coffee.
“Performance in the office is what matters,” Cartwright says. “We’re working on cutting edge data science here, and these women are going to take these skills and go on to start companies and build teams.”
