When Vienna H entered her first AI4ALL class, she had never directly engaged with artificial intelligence (AI). Over the course of the program, she went from viewing it as a nebulous concept—a “new and emerging” technology with little bearing on her future or present—to implementing it to find alternative approaches to predicting COVID-19 based on analysis of electronic medical records. This project—in addition to four others, each focused on using AI in a different way to improve COVID-19 prediction—represents the culmination of the 2020 UCSF AI4ALL program. The camp concluded on July 31st with a symposium where students presented their teams’ original research into applying artificial intelligence methodologies to COVID-19 data (symposium talks can be viewed here).
For many of the students like Vienna, the UCSF AI4ALL program marked their first engagement with AI. In just three weeks, they went from that first introduction to producing projects comparable to those of full-time researchers.
“If I didn’t know it was high school kids presenting, they could have been graduate students or postdocs easily,” said Dr. Marina Sirota about the student presentations at the end of the program symposium. Dr. Sirota is the AI4ALL Program Director and Associate Professor at the UCSF Bakar Computational Health Sciences Institute.
The projects were produced as part of the UCSF AI4ALL summer program, founded and directed by Dr. Sirota, and co-directed by Dr. Tomiko Oskotsky. The program, which is now in its second year, is linked to the broader AI4ALL organization, but is the first and only affiliated program to focus specifically on biomedical applications of AI.
The first week of the program centered on the AI4ALL core curriculum and helps students develop the basic programming and data-combing skills necessary for AI implementation; the second and third were devoted to working on the research projects. Daily faculty guest speakers, including Atul Butte, Ida Sim, and Sergio Baranzini, and panels provided perspectives on applications of AI in the biomedical field and more broadly on the experience of doing novel scientific research.
The UCSF program, supported by AI4ALL and the Bakar Computational Health Sciences Institute, in particular emphasizes the importance of hands-on, realistic research experience, aiming to give students a taste of work in a research field.
“[Research] is not a clear, straightforward path,” said Vienna, describing one of the most important lessons she learned during the program. “It’s definitely got a lot of twists and turns. So it’s okay if things don’t go the way you initially envisioned going.”
Throughout the program, students tackled the “twists and turns” with the full support of their peers and the mentors engaged in the program.
“One of the most beautiful things that I saw come out of the UCSF AI4ALL program was the connection that the students forged with like-minded individuals who were supportive of each other,” Oskotsky said.
Jessica S, another of this year’s students, remembers logging into the Zoom classroom for the first time and seeing that all of the members were non-white women. For her, the realization that she would be learning with a diverse, supportive peer group, so different from her previous computer science classes that were “mostly white boys,” was “really exciting.”
AI4ALL as an organization aims to increase diversity in AI, with each individual site providing AI education to a slightly different subset of underrepresented communities in the field.
“AI4ALL recognizes [the importance of multiple perspectives],” Oskotsky said. “It invites and encourages people from different backgrounds to come in to learn about AI and to go forth and be leaders in this—to not just lead the studies, but also establish policies to lead others in this work.”
In order to facilitate program participants’ engagement in these broader discussions, AI4ALL maintains an alumni network known as “Changemakers.” The network is a place where AI4ALL alumnae, called ‘Changemakers in AI,’ are able to continue learning about AI and computer science techniques. It also provides opportunities to engage with the AI research community far beyond their initial program experience through mentorships, invitations to attend or speak at conferences, and research opportunities.
“[The Changemakers network] is a really great way to stay involved in the AI4ALL community, but also to increase the visibility of all the changemakers and the access to AI education,” said Isha K, a member of the inaugural UCSF AI4ALL class who returned as a TA for this year’s program.
The connections formed through the Changemakers network allow students to continue learning AI methodologies and discovering more about the future of artificial intelligence. That forward-facing outlook is something that some students see as the most important aspect of the program—and a reason prospective students ought to apply.
“I can really see [AI] playing an integral part in our next generation, because it’s such a rapidly advancing thing. I’ve gained a much more rounded understanding of what AI is, and I understand a lot better how it’s going to be used,” Vienna said. “Getting to know what the future will look like is going to be a great start. AI4ALL really introduces you to the amazing power you can harness with artificial intelligence.
“I would say, just go for it. It’s a great program to learn more about what the future could potentially look like and what the future holds.”