Shouldn't the Starbucks app learn who pays with the app, and land them on the pay tab when they open the app while in a Starbucks? Shouldn't Instagram learn and auto-populate tags people use frequently? Shouldn't medical supply systems likewise learn what clinicians often order? So much attention gets focused on cutting-edge applications of artificial intelligence (AI), like self-driving cars or robotic surgery, that many mundane but immediately useful time- and cost-saving applications get overlooked, says John Zimmerman, professor of Human-Computer Interaction at Carnegie Mellon University and a 2020-21 AAAS Leshner Leadership Institute Public Engagement Fellow.
During his fellowship, Zimmerman has been talking with different groups, often user-experience (UX) designers working in technology, healthcare, education, and more, about how to “get their money’s worth” out of AI. Often, the design innovations (i.e., innovations in how something is used, not in the technology itself) Zimmerman encourages can be implemented by a company’s IT staff without much if any external support.
“I want to help UX designers envision new things where we are highly confident AI can do these things,” he explains. These kinds of uses might be, for example, generating insights for doctors to consider – as opposed to diagnosing a patient for the doctor, a more difficult and potentially contentious use. Zimmerman has also been engaging with AI developers about ways to work more closely with domain experts and end users as they formulate ideas for research and development.
As he has continued to have conversations around this topic, Zimmerman has had the chance to refine his message, a key component of the AAAS science communication training he received at the start of the fellowship. He’s found this to be a useful exercise in iterating on a message, finding stories and framings that get his point across most effectively, like the Starbucks and Instagram examples.
Another of Zimmerman’s public engagement goals is to engage local high school students in the Pittsburg area to help them see themselves in AI careers. Such engagement is a means to increasing diversity in AI – including not just demographic diversity, but also diversity of interests and approaches to problems. His son attends an arts-focused high school, so he may build on this connection in the future, since the coronavirus pandemic has thus far slowed his progress toward this goal. In the meantime, he participated in a faculty committee to bring in a vice president focused on diversity, equity, and inclusion, and has been working on ways to improve their recruitment and retention of more diverse faculty.
“The Leshner Leadership Institute fellowship has changed my perspective a lot,” Zimmerman says. "The conversations with other fellows, the group meetings, the training week, they were really helpful in having a broader view of what outreach is.”
The fellowship experience also significantly changed how he writes grant proposals. Since the fellowship orientation last June, Zimmerman says he has begun incorporating engagement in increasingly bigger ways as formal parts of his grants, building it into the essential research activities instead of something done on the side. In a recent NSF grant, he proposed giving workshops to professional designers at South by Southwest on how to design AI products and services. He also talks more to his PhD students more about engagement. “I think it makes our proposals more competitive,” he adds.
The AAAS Leshner Leadership Institute was founded in 2015 and operates through philanthropic gifts in honor of CEO Emeritus Alan I. Leshner. Each year the Institute provides public engagement training and support to 10-15 mid-career scientists from an area of research at the nexus of science and society