Presented by Capital One

At Transform’s second annual Women in AI Breakfast presented by Capital One and Intel, powerful women from tech companies across the industry gathered to talk about how women, particularly women of color, can take their seats at the table in the technology industry.

“Much has been written about the industry’s pipeline problem, and how we can increase diversity in tech companies,” said Carla Saavedra Kochalski, director of conversational AI and messaging products at Capital One who provided opening remarks. “Many say they don’t hire women or people of color candidates because there aren’t enough qualified candidates. And it’s true — there are fewer of women than men with computer science degrees.”

But what if that low number is more of a symptom, than a cause? Roughly 56% of women leave the technology industry by the mid-level point in their careers. These women are often driven out of a promising career in tech by a number of factors, including institutional barriers, microaggressions in the workplace, and bias. But what about the 44% who stay in tech?

Women who receive regular training are more likely to stay in their tech careers than those who don’t. They succeed because they get regular feedback, both positive and negative, and tend to have stronger role models and peer networks.“

A panel discussion followed, with Kay Firth-Butterfield, Head of AI and Machine Learning and Member of the Executive Committee, World Economic Forum; Dr. Timnit Gebru, Co-lead of Ethical AI Research Team, Google Brain; Francesca Rossi, IBM Fellow and AI Ethics Global Leader, IBM Research; and moderator Jaime Fitzgibbon, Founder & CEO, Insights.

“We have a long way to go, which is why role models are very important,” Rossi said. “Women need to understand much earlier, that they can have place in a technological environment.”

Firth-Butterfield noted how essential it is to bring women with different cultural and academic backgrounds into the AI and tech space.

“Even with a lot of support behind it, it’s very difficult to move women into the STEM careers,” she said. “One of the things that we are doing at the forum is creating an AI Youth Council to bring young people in from around the world to talk about AI and the issues that affect them today.”

Gebru emphasized that the term “underrepresented majority” is important to bring forward because it highlights the fact that most of the groups talked about are not necessarily a minority in the world, but the majority. These movements have to have an intersectional lens, take into account the multiple angles and dimensions of marginalization, and keep the most marginalized of these groups front and center, she said.

“In these discussions, it’s essential to look at intersections of race, class, gender, and all of these different dimensions of marginalization, or I think that we will fail to advance all women,” Gebru said. “I think that’s true in anything, and especially in AI.”

But in these conversations about diversity and inclusion, it’s easy to fall into the trap of talking on behalf of — and over — underrepresented majorities.

“There’s the term ‘nothing about us without us,’” Gebru said. “When you’re advocating for a certain group of people, are you lifting their voice, are you passing the mic, are they talking, or are you talking on their behalf? Advocacy starts from watching out for our own actions and behavior.”

An important part of the solution is self-education, she added, which allows you to view the problem in a more holistic lens. She also offered tools for the quest toward equality and recommended attendees follow folks like Meredith Broussard, AI researcher and author of “Artificial Unintelligence: How Computers Misunderstand the World”; Ruha Benjamin, Founding Director of the Ida B. Wells Just Data Lab and author of Race after Technology; Simone Browne, author of Dark Matters; and other women of color who are advancing the conversation around AI ethics and race.

It takes establishing initiatives that support and uplift women in the workplace, said Huma Abidi, Senior Director of AI Products, Intel, who founded the Women in Machine Learning forum at Intel. The company was able to meet its goal for 2020, which was to represent women at all levels. The next goal is to have 40% representation of women in technology at all levels of the company by 2030.

“More and more women are stepping up to address this challenge to correct bias in AI and to ensure that AI solutions are ethical and fair,” Abidi adds. “I get inspired, when I see Fei-Fei Li leading human-centered AI at Stanford, when I hear about Meredith Whittaker working on the social implications of AI at the AI Now Institute, and more. We have a lot of work ahead of us, but I see that we are heading in the right direction.”