While studying cancer biology as a health sciences student at McMaster University in 2016, Andrew Leber started to wonder how artificial intelligence might help diagnose and improve cancer treatments. He brought together 10 friends, also science students, for a reading group focused on technical concepts in machine learning.
But it turned out many more students were interested. Leber and friends opened the reading group to a wider audience, and within a month it had 50 members. A few months later, Leber launched the McMaster AI Society, which blossomed into one of McMaster University’s largest student-run clubs. The group received a sponsorship from Microsoft and has since grown to more than 1,000 members, many of whom are from faculties such as business, the humanities and social sciences.
One of them is Sarah Bowron, a third-year sociology student, who was concerned about the legal landscape and privacy implications of AI: “What are the privacy boundaries? How is [AI] impacting people in ways that we really don’t know about?”
Early on, the club’s popularity caught the attention of an engineering faculty member, who floated the idea of creating a new undergraduate course. Leber led the committee to develop the course, and Innovate 1Z03 is now offered through the university’s innovation minor program, a collaboration between the faculty of engineering and the DeGroote School of Business. It’s a bit of a departure from typical AI courses that are available through computer science and engineering departments in Canadian universities, which tend to teach algorithmic techniques to program and analyze AI rather than the societal implications. Instead, Innovate 1Z03, and other new interdisciplinary courses at the University of Toronto, are teaching AI as both an art and a science.
“Students are going into a world where they hear about AI all the time,” says Matthew Jordan, the professor who teaches Innovate 1Z03. “This new technology is going to be a fixture of their life, and it’s over-hyped and misunderstood.” Throughout the semester, Jordan’s students reflect on how AI is presented in news headlines and wrestle with questions such as: how close is AI to reaching human-level intelligence? Is AI replacing the need for human labour? Is AI in fact a new technology? (Spoiler alert: the field has existed for 75 years.)
For one assignment, students assume the role of a CEO of a global AI company and work in teams on a roadmap for the AI industry or to develop a new AI application. Akil Hamilton, a fourth-year software engineering student, designed a virtual assistant for physicians to offer customized treatment recommendations based on a patient’s experience. Hamilton says a problem emerges when AI algorithms are shown or “trained on” data from an overrepresented demographic group. For example, if an image recognition algorithm is trained on more light-skinned faces than dark-skinned faces, the AI would be better at detecting light-skinned faces. “If we are looking to give an accurate health-based recommendation, [this kind of algorithmic bias] is something we will need to overcome,” Hamilton says.
RELATED: Using traditional Inuit knowledge and Western science to study Arctic marine life
At the University of Toronto, Karina Vold teaches two courses that encourage students to think about AI like philosophers. In The Limits of Machine Intelligence, students unpack hidden assumptions about the concept of “intelligence,” stepping back to reflect on what it means to consider something “conscious” or “intelligent” in the first place. “In some ways, it’s a dream-shattering course,” says Vold, because there’s an idea that AI has the potential to exhibit human-level thinking or self-awareness. But in Vold’s course, the takeaway is often, “Hey, AI is not there yet. But here’s why,” she says.
In many ways, these new interdisciplinary AI courses are both sobering and illuminating, putting in the foreground the importance of deep, critical thinking, a long-valued skill in the liberal arts. “It would be a harm for us as a society to let a technology be built that has such a widespread impact without having some critical reflection on what that impact is, and without trying to really understand it,” Vold says.
This article appears in print in the 2022 University Rankings issue of Maclean’s magazine with the headline, “Machines are us?”