Diversity in AI Is Essential to Our Future

Right now, many of us are thinking about the ways AI will transform our daily lives. Some of the prospects are exciting—like early disease detection. Others are alarming—like wrongful arrests due to bias in facial recognition software. In either case, the stakes are high, so it’s important that the people who are designing and developing this technology are doing so thoughtfully and with a deep understanding of its human impacts.

As Dr. Fei-Fei Li—a renowned AI researcher and author of The Worlds I See: Curiosity, Exploration and Discovery at the Dawn of AI—recently put it, “Even though it's called ‘artificial intelligence,’ it's made by people. It's built by people. It's deployed by people.” It matters, then, who those people are—and whether they are bringing a wide range of perspectives and experiences to their work.

Today, AI is still a relatively homogenous field—dominated by a handful of predominantly white male technical experts working at a handful of elite institutions, recruited from the same handful of elite universities. If they’re the ones teaching AI how to think and understand the world, the technology they create is likely to replicate their biases and blind spots, making that technology less effective and our society less equal.

Pivotal Ventures is committed to advancing social progress and expanding women’s power and influence in the United States. And because AI is going to play such a powerful role in shaping how we all live and work, we think it’s crucial that the field is equitable and inclusive. That’s why Pivotal has been working since 2017 to ensure that women from diverse backgrounds are involved in the design and development of the technology.

Telling students “AI needs your perspective”

One challenge to diversity in AI is getting the field on more students’ radars, empowering those students to imagine themselves in the field and see AI as a tool to help them solve the problems they care about. AI4ALL, a Pivotal partner co-founded by Fei-Fei Li, provides opportunities for college students of all backgrounds to learn about the field and better understand how they can contribute to it.

AI4ALL’s active alumni network is filled with changemakers who are applying AI technology to a wide range of important problems—improving surgery, increasing access to emergency care, and reforming the criminal justice system, for example. And just as importantly, they’re proving that when people with different perspectives gain access to these powerful tools, they can use them to drive progress on problems that others may have overlooked.

Helping students show their potential

Another challenge to diversity in AI is the fact that employers tend to recruit from only a small handful of elite schools, making it difficult for other students to break into the field. The question, then, as the founder of Pivotal partner Break Through Tech, Dr. Judith Spitz, explained in Forbes, becomes “How can a student demonstrate their potential, if they lack the access that comes with privilege?”

Break Through’s answer is to annually connect students from approximately 120 colleges and universities—most of whom are women of color, and many of whom are the first in their families to attend college—to an intensive summer introduction course to AI and Machine Learning hosted by faculty and graduate students at Cornell Tech, MIT and UCLA.

Next, the students participate in AI Studio challenge projects, giving them the hands-on experience they need to land internships. Then, interning at organizations like Microsoft, NASA, Goldman Sachs, and Google, they get to solve real-world business challenges, demonstrate their skills, and position themselves for fulltime job offers upon graduation. By 2024, the program will serve 1,500 students annually, many of whom will go on to work in AI, and, in turn, shape the future of the field.

Creating new pathways to Ph.D. programs

Another group that will have an outsized impact on the future of AI is the small number of people graduating with Ph.D.’s in computer science—only about 2,600 each year in the United States. That credential helps position the people who hold it to shape the research agenda and lead the development of the technology for decades to come.

The challenge is that there are very few spots in computer science Ph.D. programs, so admission is very competitive—and it’s very difficult to get into one without a robust research portfolio as part of your application. But research opportunities are limited and go disproportionately to students with built-in networks at the elite institutions that dominate the field, making it hard for other students to compete against them.

Our partner Computing Research Association opens up pathways for undergraduate women in computer science to conduct technical research, giving them a fair shot at admission to those highly competitive Ph.D. programs and the chance to contribute to the future of AI at the highest levels of the field.

Demanding accountability

There have been many examples of what AI-powered bias looks like. One of the most well-known is AI researcher Dr. Joy Buolamwini’s 2018 finding that facial recognition software misidentified darker-skinned females 35% of the time, compared to less than 1% for lighter-skinned males. That research, called Gender Shades, informed her popular TED Talk, How I’m Fighting Bias in Algorithms, in which she called for the need to create AI teams populated with diverse individuals “who can check each other’s blind spots.” As a result of her work, companies including Amazon, IBM, and Microsoft have agreed not to sell facial recognition software to law enforcement institutions until there is governance in place to prevent bias.

Dr. Buolamwini went on to found the Algorithmic Justice League (AJL), a Pivotal partner. AJL is dedicated to expanding public awareness of the risk of bias in AI and holding creators of AI accountable when their applications do harm to people in the real world.

As Dr. Buolamwini has said, "Whether AI will help us reach our aspirations or reinforce the unjust inequalities is, ultimately, up to us." Pivotal Ventures is proud to stand behind so many changemakers who are taking that responsibility seriously.

Dr. Fei-Fei Li’s book, The Worlds I See: Curiosity, Exploration and Discovery at the Dawn of AI, was published by Moment of Lift Books, an imprint created by our founder, Melinda French Gates, in partnership with Flatiron Books.