There's a Massive Gender Gap in AI, but Tech Education Programs for Young Girls Aim to Close It Iridescent, a tech education nonprofit, teaches women and girls about AI, tech entrepreneurship and more. So far, it's trained more than 114,000 people from 115 countries.

Tara Chklovski had the weight of her family on her shoulders.

Growing up in India in what she calls a lower-middle-class family, it was difficult not to notice the poverty around her -- especially the pivotal role chance plays in determining the family someone is born into, as well as their access to education, healthcare and opportunity.

Chklovski's parents encouraged her to pursue a subject that intrinsically advanced the world, hoping it would also advance their family's situation. "That is definitely the mantra of the times in India," says Chklovski. "The lower middle class has this drive to get a degree in engineering or medicine or technology so you can lift your family out of poverty."

She came to the U.S. to study aerospace engineering and quickly found that that same drive to pursue tech didn't apply -- especially in the women she met. The U.S. may be a more developed country, she thought, but here, women actively kind of closed doors to their potential, making blanket statements like, "I'm not good at math." Chklovski was stunned someone could say that with a straight face, but she was also intrigued. She did some digging and concluded that the "huge, transforming lever" that is education was at the root of the problem.

Chklovski wanted to become a pioneer in the area, so she left her PhD program to start Iridescent, an educational nonprofit that says it's helped train more than 114,000 people from 115 countries since its 2006 launch. The organization sets its sights on empowering young girls and mothers to become tech leaders in communities across the globe, partnering with the likes of Google, GM and Boeing in its mission to teach AI and entrepreneurship to people who have identified problems they'd like to solve in their own communities.

It's a man's, man's, man's world -- especially if you're looking at the technology sector.

At eight major tech companies, women make up between 27 percent (Microsoft) and 47 percent (Netflix) of the workforce, according to data recently compiled by Statista. But when it comes to tech jobs in tech -- think developers, engineers and the like -- the numbers are bleaker: At seven tech giants with data for that category, an average of eight in 10 jobs belonged to men.

Research from longer-term studies backs that up. One report analyzing the number of women in selected STEM fields from 1990 to 2013 found that the number of women in computing fell from 35 percent to 26 percent over that time period. Break it down by race and ethnicity, and you'll find that black women hold just three percent of jobs in computing, while the statistic is one percent for Hispanic women and a fraction of a percent for American Indian and Alaska Native women.

In AI, the problem is no different. Just 18 percent of authors at leading AI conferences are women, and more than 80 percent of AI professors are men. Women make up 15 percent of AI research staff at Facebook and 10 percent at Google, according to a new report from the AI Now Institute, and there is no public data on trans workers or other gender minorities. Another key point the report makes: Focusing solely on improving the ratio of women in tech is "too narrow and likely to privilege white women over others."

"To date, the diversity problems of the AI industry and the issues of bias in the systems it builds have tended to be considered separately," the report reads. "But we suggest that these are two versions of the same problem: Issues of discrimination in the workforce and in system building are deeply intertwined."

Indeed, the issue not only impacts women and girls who could be contributing to the field but also the broader public -- the people the tools are being made for.

"This is no longer about building a fun toy or some sort of technology that has tangential impact," says Rumman Chowdhury, who leads Accenture's Responsible AI initiative. "[These are] technologies that make decisions about our lives, that shape our very existence, and it's very problematic not to have representation in the room -- whether across gender, sexuality, even things like geography. You build better products when you have a diversity of people, a diversity of voices [in] what's being made. One thing Silicon Valley is learning is we're not just building toys for people in Silicon Valley now -- we're building things that impact all kinds of human beings, and for that you do need representation."

AI programs designed specifically for girls and women aim to make a change.

One of these is Iridescent's AI Family Challenge, a series of mentor-run workshops for families across the globe, centered around empowering mothers and young girls. Families identify a community-centric issue they'd like to help solve, then learn to use AI -- specifically, a platform called "Machine Learning for Kids" that runs on IBM's Watson -- as a tool to help get there. It helps walk users through training a model to recognize patterns, such as specific types of images or even emotions. One example: After working with a psychologist to create training data, a mother-and-daughter team from Palestine created an image-recognition system analyzing children's drawings to gauge whether a child is feeling depressed or experiencing domestic violence. In May, Iridescent will host the final round of its first AI Family Challenge World Championship, where families from Palestine, Pakistan, Spain, Bolivia, Uzbekistan and the U.S. will present their ideas to a panel of judges in Silicon Valley.

"The AI Family Challenge alumni are really mothers and women who are from low-income communities but have been… introduced to technology, and they are really, really ripe for being connected to accelerators and small business [or] startup enterprises," says Chklovski. "These are extremely inventive uses of AI from the field right now, from all across the world, from very, very unlikely sources… That's the power of education. We just need to give them the right tools."

In five years, Chklovski predicts industry-level AI tools will be accessible to the broader public, but she believes one vital ingredient will still need nurturing in women and girls: Courage to believe they can build something that will make a change in their communities.

"That sense of empowerment still needs to be there -- and that's our goal," she says.

That idea is spotlighted by testimonials from mothers across the globe. Rustamkhuja, a mother from Uzbekistan who participated in Iridescent's AI Family Challenge, said, "We gained knowledge on robotics [and] AI. We made new friends and brainstormed an idea of how to use AI to help improve our lives." A mother in Nigeria, Laura, said, "Our analytical thinking has improved, and we now explore different ways of solving or arriving at a conclusion. We also realize that we should always keep trying with projects -- not getting it right the first time does not mean we should give up."

Why is the tech industry, including the AI field, so short on women? One key reason, experts say, is that since many girls aren't encouraged to pursue STEM, their interest in the field fades. In fact, a survey of young women in Europe commissioned by Microsoft found that their interest in STEM peaked around age 11 or 12 and dropped off sharply between ages 15 and 16. Another study of 6,000 girls and women in the U.S. found that about one-third of middle school girls thought they might pursue a career in coding, but just 27 percent of young women ages 18 to 30 said the same.

There's a light at the end of the tunnel, however: Things seem to change when young women have female role models in tech. Middle and high school girls who personally knew a woman in STEM were about 17 percent more likely to feel powerful working on STEM activities. And middle and high school girls who participate in STEM clubs or activities were 26 percent more likely to say they felt powerful doing STEM activities.

The solution here isn't just about encouraging girls receive to pursue STEM and AI; there's also a conversation to be had about supporting women throughout their careers in the field. Almost half the women who enter the field of technology eventually depart -- more than double the percentage of men who do the same, according to a report from the National Center for Women and Information Technology.

"There is a lot of focus on girls getting interested in STEM, but we have a very immediate problem to address about women who are in STEM today being protected, promoted and retained by companies," says Chowdhury. "While it is a noble endeavor to work on the people who will be in the workforce 10, 15 or 20 years in the future, look at the people who are there now. How are you protecting and serving them?"

Iridescent isn't alone in its mission to bring AI education to girls around the world.

There's iD Tech, a series of summer camps across the U.S. that currently offers an all-girls robotics course called AI Lab. GirlsComputingLeague, a nonprofit aiming to empower underrepresented groups in tech, hosts an annual AI Summit in Virginia for middle and high school students interested in the field. And in Singapore, the nonprofit 21C GIRLS offers AI education programs for local girls, including a three-month course on machine learning.

Last year, Women in AI (WAI) offered its first summer camp to teach young girls about AI and robotics, and Microsoft offered an AI bootcamp for girls in STEM, including a hackathon day where girls worked in teams and used machine learning to predict chances of breast cancer and create a bot service with medical FAQs to support doctors.

Girl 2.0, a California-based nonprofit that aims to close the tech gender gap by providing free computer science education, hosts regular AI workshops for girls, while GAITEway, an acronym for Girls' Artificial Intelligence and Technology Education, provides AI workshops for middle school girls in the Bay Area.

Then there's AI4ALL, which offers summer camps at colleges across the country to teach key facets of AI to underrepresented high school students. At Stanford University, the program is specifically aimed at ninth-grade girls from a range of financial and cultural backgrounds, and the annual three-week residential summer program features lectures, field trips and meetings with potential mentors. Hands-on projects from past programs include applying natural-language processing to disaster aid, using machine-learning algorithms to comb the human genome for signs of cancer and exercises in programming autonomous vehicles.

Amy Jin, an inaugural program participant in 2015, heard about AI4ALL through her high school's Women in STEM club. "For my group project, I worked on a computer vision-based hand hygiene monitoring system to combat hospital-acquired infections," she said through a program representative. "At Stanford AI4ALL, I wasn't just exposed to the power of technology, but also to the idea that leveraging the power of technology for social good was at my fingertips."

Jin said she was also struck by the sheer variety of AI's possible real-world applications -- for example, how natural language processing can be used to mine tweets for disaster relief or how bioinformatics could lead to more information on cancer genetics.

"AI will change the world," the AI4ALL website reads. "Who will change AI?"