More women and minorities must work in tech, or else they risk being left behind in every industry.
This grim future was painted by Artificial Intelligence (AI) equality experts who spoke at a conference Thursday hosted by LivePerson, an AI company that connects brands and consumers.
In that future, if AI goes unchecked, workplaces will be completely homogenous, hiring only white, nondisabled men.
“In this bleak depiction of our future, decades of fights for civil rights and equality have been unwritten in a few lines of code,” said EqualAI executive director Miriam Vogel at the conference in Brooklyn, N.Y., called “Boundary Breakers: Women Driving The Future of Tech.”
Women and minorities are not building AI, and therefore, they are not being represented in popular algorithm-based products, according to Vogel. Algorithms find patterns throughout history, and essentially make predictions based on stereotyping, she said. Already, facial recognition technology perceives white men more easily than women and minority features, Vogel said. Minorities are being denied credit cards due to where they live [and] due to their area’s history of poor credit.
Large companies are using AI in their hiring practices, which leads to inevitable implicit bias that is difficult to detect, Vogel said to the crowd of about 55 tech thinkers.
“The people employing these AI products probably don’t even know their outcomes are discriminatory,” said Vogel. “AI becomes a mirror. It reflects and it magnifies the biases in our society.”
Guest speaker Cathy O’Neil, who authored “Weapons of Math Destruction,” explained how hiring bias works with AI: company algorithms are created by (mostly white male) data scientists, and they are based on the company’s historic wins. If a CEO is specifically looking for hirees who won’t leave the company after a year, for example, he might turn to AI to look for candidates based on his company’s retention rates. Chances are, most of his company’s historic wins only include white men, said O’Neil.
“Algorithms are presented to us as objective truth that will solve problems that seem unsolvable. But what we really have is people who have their definition of success who are hiding behind the authority of mathematics,” O'Neil said. “The truth is, it’s not that mathematically complicated...It’s about pattern matching, and the presumption that what worked before will work again. And guess what, we don’t actually want to repeat history, do we?”
To fight the rise of bias in AI, more representation is critical in the computing workforce, where only 26 percent of workers are women, 3 percent are African-American women, and 2 percent are Latinx.
Speaker Judith Spitz said only around 18 percent of computer science graduates are female, and that this number has been static since 2008. Spitz built an internship program called Women in Technology and Entrepreneurship in New York (WiTNY), which finds paid computer science internships for low income, female high school freshman and sophomores.
“The confidence they got from having met with executives, the experience that they’d been there, and project they worked on...I can tell you, the confidence they get from one paid summer tech internship, they’re off and running...One of our students is sitting on five offers this summer. It’s really transformational.”
Panelist Lisa Kay Davis, a journalist and social strategist at IBM, credits her mentors and her ability to speak up for her rise.
“I’ve been groomed to be the type of person who speaks up and advocates, and ask ‘is this really true?’” said Davis. “Especially in tech, you get to ask that question all the time. I like to push people...there are always ways to create opportunities for other women.”
Davis encouraged women to negotiate for their worth.
“Be greedy about getting access to information,” said Davis. “...I relish sharing that information and asking ‘how much do you make?’ It makes some people uncomfortable, but you have to know. Sometimes there is a $20,000 gap between what you’re making and what the person next to you is making, and you’re doing the same job”
For Vogel, closing the gap between men and women in tech is not just a moral imperative. It could curb economic disaster for women and minorities. She said she was alarmed that her two daughters were not learning to code in middle school.
“I’m used to being the only woman in the room...but I thought it would be different for my two daughters,” she said. We all need to ask if our daughters, nieces and neighbors are being exposed to the modern language of coding, computational thinking, and data literacy.”