Inaugural SRI Faculty Fellows build bridges between disciplines and forge new areas of research
With their landmark donation of $100 million to the University of Toronto in 2019, philanthropists and business leaders Gerry Schwartz and Heather Reisman initiated a transformative project that has forged vital connections across the university and beyond. The gift initiated both the construction of the Schwartz Reisman Innovation Campus—a 750,000 square-foot complex set to become Canada’s biggest university-based innovation node—and the Schwartz Reisman Institute for Technology and Society (SRI), placing U of T as a leader at the forefront of Toronto’s tech boom.
The mission of SRI is to deepen understanding of powerful technologies such as artificial intelligence (AI) through interdisciplinary research, and develop human-centred solutions to ensure new technologies improve life for everyone. Among the institute’s key mandates is a multifaceted research program that draws upon U of T’s rich community of scholars, with the goal of developing new approaches towards research, enhancing interdisciplinarity, unlocking innovative solutions to regulatory challenges, and promoting the use of AI for social good.
As part of its research program, SRI offers annual Schwartz Reisman Fellowships to support projects from U of T faculty and students from across the sciences, social sciences, and humanities that engage with pressing issues at the intersection of technology and society, including the four “conversations” that guide the work of the institute. Schwartz Reisman Fellowships support researchers looking to build connections between existing fields of inquiry, drawing upon the goals outlined in SRI’s 2021–24 Strategic Plan.
In 2020, SRI granted its inaugural Schwartz Reisman Faculty Fellowships to four remarkable U of T faculty members: Kristen Bos, of the Department of Historical Studies and Women and Gender Studies Institute; Aleksandar Nikolov, of the Department of Computer Science; Nisarg Shah, of the Department of Computer Science; and Karina Vold, of the Department of Philosophy and the Institute for the History & Philosophy of Science & Technology.
Despite the considerable challenges of the COVID-19 pandemic—which resulted in the university working remotely for more than 18 months, and rendered in-person meetings, research, and events challenging—the inaugural cohort of SRI Faculty Fellows found ways to carry forward, building bridges between disciplines and forging new areas of research, as well as engaging with SRI’s community to share their ideas and promote new connections.
Kristen Bos: Visualizing environmental justice through data
In addition to teaching Indigenous science and technology studies, Kristen Bos is the co-director of U of T’s Technoscience Research Unit and leads the Environmental Data Justice (EDJ) Lab, which focuses on the relationships between data, pollution, and colonialism.
In her fellowship project, “Expanding Environmental Data Justice and Digital Modalities,” Bos investigated how environmental data can be used to better serve the communities who live amidst intensive pollution corridors such as Canada’s Chemical Valley.
“SRI’s fellowship went towards funding two additional community research positions for Anishinaabekwe Vanessa Gray and Beze Gray, whose time was spent combining a critical analysis of the limits of current pollution data with practical ways of making problematic data more useful to community members,” Bos explains. “Through their work, we were able to finalize a project with the National Pollution Release Inventory (NPRI), which will address our key research challenge of how to make NPRI’s inadequate source of industry-reported data more meaningful for community members.” Significantly, Bos negotiated a release from the standard Crown copyright over research materials—a win for Indigenous data sovereignty.
Bos is currently developing a three-part visual series contextualizing pollution in Ontario’s oil and gas extraction sector, including petroleum refineries, petrochemical plants, and plastics manufacturing. Her project was recently profiled by U of T’s Groundbreakers video series.
Aleksandar Nikolov: New approaches to privacy
Aleksandar (Sasho) Nikolov’s fellowship project focused on a new approach towards data privacy known as differential privacy, which he observes is “increasingly seen as the gold standard of rigorous privacy definitions.” The benefit of differential privacy is that, in the event of a security breach of a given data set, it guarantees that even a well-informed adversary will not be able to tell with confidence if a particular individual’s data was used.
“Most privacy regulations are phrased in terms of the ability to identify a specific individual from information published about them,” Nikolov explains. “Differential privacy, by contrast, guarantees that the risks posed to any individual are not much greater when they contribute information to a data set, compared to when they do not contribute any data.”
In a SRI blog post co-authored with SRI Faculty Affiliate Nicolas Papernot, Nikolov considers the implications of a differential privacy approach for privacy law, proposing that “meaningful privacy regulation [should] focus on properties of the algorithms analyzing the data, rather than on the anonymity of the data.”
Among the goals of Nikolov’s research is addressing gaps in the conceptualization of privacy between computer science and legal scholarship. Computer scientists may build privacy protocols for systems they develop without being fully informed about the definition of privacy in a legal sense, or how privacy laws might vary in different territories.
To bridge this disciplinary gap, Nikolov collaborated with Papernot and SRI Research Leads Lisa Austin (Faculty of Law) and David Lie (Department of Electrical and Computer Engineering), to co-author a blog post on mismatches between regulation and data analytics, and propose new definitions for a revised regulatory framework. This text was expanded into an article presented at the 2021 Privacy Law Scholars Conference, has been the subject of recent SRI Seminars by Nikolov and Austin, and remains an ongoing project.
Collaborating with experts on privacy outside the domain of computer science was central to the success of his fellowship, Nikolov observes: “This project allowed me to reach a different audience, including privacy law scholars and regulators. In the course of our collaboration, we had a meeting with a Member of Parliament, and shared our analysis of the shortcomings of Bill C-11.”
Nisarg Shah: Voting and algorithmic fairness
Nisarg Shah’s fellowship explored designing fair and efficient AI technologies to be used for participatory budgeting. “Participatory budgeting is a process by which residents of a geographical region vote on the allocation of public budget to infrastructure projects in their neighbourhoods,” Shah explains. “This process underlies the annual allocation of hundreds of millions of dollars worldwide.”
As part of his research, Shah developed new design approaches for balloting, allocation rules, and fairness considerations. The fellowship enabled Shah to conduct user studies to evaluate proposed methods in terms of fairness and efficiency, enlist postdoctoral fellow support, and aided Shah in publishing 13 papers and four pre-prints on topics relating to voting and algorithmic fairness, including a collaboration with Nikolov on data analytics.
During the course of his fellowship, Shah delivered 16 presentations to a wide range of audiences, including the Canada Revenue Agency, the Ontario Centre of Innovation, the Regional Municipality of York, Toronto Public Library, the Centre for Advancing Responsible and Ethical Artificial Intelligence (CARE-AI) at the University of Guelph, the Vector Institute, and SRI’s Seminar Series.
Shah’s work on participatory budgeting marks an exciting extension of his innovative work in computational social choice, which includes the websites RoboVote.org and Spliddit.org—two not-for-profit platforms that have been employed by more than 200,000 users to date. His successes have led to IEEE Intelligent Systems recently naming him one of ten outstanding young scholars to watch.
Karina Vold: How humans can learn from machines
Karina Vold’s fellowship project focused on human and machine learning across different domains, asking how we can use AI systems as “scaffolding for human learning” to help us acquire new knowledge. A philosopher who explores the intersections between AI and cognitive science, Vold’s work is concerned with the ethical and social implications of new tools, and what technology can teach us about being human.
As part of her research, Vold has developed curriculums for new courses at U of T on AI and machine learning from a humanities perspective, on the history and philosophy of AI, and on the limits of machine intelligence, which challenges students to take a closer look at the philosophical questions surrounding the prospect of “intelligent” artificial systems.
“One of the major outputs of my SRI Faculty Fellowship was leveraging the funds for time to apply for a SSHRC Insight Development Grant aimed at growing the human and machine learning project,” observes Vold. “Some of these funds enabled me to support incoming graduate students and to hire them as RAs to support further SRI-related research.”
During her fellowship, Vold edited two special volumes, published five papers, and gave seven invited talks, including at Oxford University, University of Exeter, U of T’s BMO Lab, York University, and The Asser Institute. Vold’s research was also featured in Macleans, and a German TV documentary on AI.
Looking ahead to the future
All four members of the inaugural SRI Faculty Fellows cohort will continue within the SRI research community as Faculty Affiliates. Each has found collaborators within the network of researchers affiliated with the institute that are leading to new research projects and generative partnerships.
With the second cohort of SRI Faculty Fellows continuing their appointments through 2023, and a third cohort of fellows set to be announced later this spring, the Schwartz Reisman Fellowship program continues to expand the conversation around AI, and assist the U of T research community towards new approaches that meet challenges of today’s new technologies.