How one company is helping to diversify the A.I. field

Dr. Taniya Mishra has been an A.I. and machine learning researcher for over a decade, and in that time, the scientist saw something crucial missing in the landscape.

“I don’t see myself represented in this field that I think of as my own,” she says.

To underscore the point, an industry report published in 2019 found the A.I. field has a “diversity crisis,” with women representing less than 20% of A.I. professors. They make up only 10% of A.I. research staff at Google, for instance, and the picture is more grim for people of color: only 2.5% of the company’s entire workforce is Black.

In 2020, Dr. Mishra founded SureStart with a goal to tackle the issue, building an A.I. optimized platform for A.I. jobs training and jobs connection. Its mission is to build a more strong and sustainable pipeline to bring students from communities currently underrepresented in A.I.

“We want to see more women, we want to see more people of color, we want to see people with different lived experiences,” she says. “And the problem is that the funnel is not wide enough.”

SureStart aims to make the funnel wider. Students are introduced to the fundamentals of A.I. through a self-paced, mentor-led program before breaking into teams to build a project using A.I. technology to solve real world problems. Throughout the program, students are connected to industry experts via seminars and discussions, and at the end of the six weeks, they virtually present their projects to a global audience.

“I’m using A.I. to build a comprehensive pipeline where students can develop all of the skills that they need, and then be connected to their first roles in industry,” she says. And they don’t have to do it alone—they are actively supported by their community of peers and mentors, and assisted by A.I. “to develop the skills to be A.I. technologists of the future.”

Ericka Corral is one such future A.I. technologist. The 22-year-old student at Beloit College in Wisconsin recently took part in the MIT FutureMakers Create-a-Thon program, a collaboration between SureStart and MIT RAISE (Responsible A.I. for Social Empowerment and Education). The six-week remote program kicked off with an immersion into deep learning models like computer vision, natural language processing, and neural networks. In that time, she also got an up-close look at the industry, connecting with professionals and asking questions about what it’s like to work with A.I.

Another added bonus of the program, says Corral, is access to seminars on topics like how to use her skills as a diverse learner, and how to overcome imposter syndrome.

“I have a really bad case of imposter syndrome sometimes,” she says. “It could get a little bit intimidating when you’re the only woman in the classroom, or people talk over you all the time. But SureStart and the MIT FutureMakers program really made sure that I’m aware that I have skills. And I’m just as good as anybody else in the field.”

During her time as a SureStart trainee, Corral felt confirmation that she wants to pursue a path in artificial intelligence and machine learning research, and was recently accepted into a PhD program.

For some SureStart trainees, the opportunity is at Consumer Reports. As the chief scientific officer for the independent nonprofit organization, Dr. James Dickerson oversees the protocol and procedure development for the company’s comparative testing of goods, services, and products—items ranging from toasters to cameras and cars. Over the decades of the organization’s existence, it’s built a glut of information that offers a deep understanding of what people are thinking about in these different categories.

Recognizing this unique position, Dickerson and some of his colleagues thought they could leverage the data using A.I. to understand more about products, “as it relates to a hazard or concern…before it actually percolates up through the traditional various flow means that we’ve dealt with and thought about in the past,” he says.

But the team lacked A.I. expertise internally. “But we didn’t have enough internal A.I. expertise, in terms of just the number of people that can bounce ideas off of each other,” he says. “And very much the collaborative point of view is what we needed to make and address some of the goals that we had.”

Through his network, Dr. Dickerson was connected to Dr. Mishra, and the two immediately saw the potential in working together. “We tossed back and forth ideas on how a collaboration could be formed between Consumer Reports as this nonprofit organization trying to do good for all in society, and Taniya’s really amazing efforts to raise the profile of those who are often not seen and not supported, who are very interested and talented in this burgeoning area of machine learning and A.I.”

Since starting the collaboration, Dickerson has worked with three SureStart trainees, including one high school student last summer, who, he says, “literally blew us away. He performed better than other consultants that we had that had advanced degrees, masters and PhDs that we’ve had in the past.”

Dickerson sees increasing diversity in the field of A.I. as crucial to its success now and in the future.

“I’m looking for diversity because it enriches the thought and enriches the creativity and enriches, ultimately, how it can be applied to the consumer marketplace,” he says. “The consumer marketplace is diverse. So those who exercise their energy to work on different projects, most certainly should come from a diverse community because our community locally, regionally, nationally, internationally, globally, is diverse.”

Mishra theorizes that when it comes to companies currently lacking a diverse workforce, particularly in the A.I. field, part of the problem is STEM pipelines, as well as the pervasive myth that to work in the field, a candidate needs a PhD. While she has the degree, Mishra regularly encounters parts of the job that don’t require one. Plus, the program is seeing students from backgrounds beyond computer science and electrical engineering, including music, theater, and philosophy.  “So it’s a little bit of myth clarification, and a little bit of widening the funnel much earlier,” she says.

SureStart is giving these students curious about A.I. the opportunity to learn and engage with the discipline in a new way, and it’s just the beginning, Mishra says, describing herself as an “A.I. optimist.”

“I see a push towards creating more inclusive, more responsible, and more equitable technology,” she says. “And for that, we need A.I. to be built by people who are more representative of the market that we hope to serve.”

Never miss a story: Follow your favorite topics and authors to get a personalized email with the journalism that matters most to you.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward