By: Kathleen Sandusky
3 Aug, 2023
The Government of Canada is currently negotiating a new health agreement with our provinces and territories that is expected to ensure interoperability of electronic health records to deliver better health care to Canadians. A timely new report published today by CIFAR through the Pan-Canadian AI Strategy assists in these efforts by providing expert guidance for policy-makers, regulators and service providers on an emerging type of machine learning that can allow multiple jurisdictions to collaborate on model training for big data health research, without sharing data.
Authored by researchers at Simon Fraser University and the University of British Columbia, Towards a Proportionate and Risk-Based Approach to Federated Data Access in Canada outlines basic steps for policymakers on how to plan, develop and implement federated learning systems.
Federated learning is a machine learning technique that holds great promise in advancing the development of health technologies for the benefit of Canadian patients, Canadian health systems and Canada’s health innovation ecosystem. It allows multiple parties, such as different health care systems, to collaborate on model training without sharing data or pooling data into a central repository.
“Bringing together Canada’s complex tapestry of single-payer health systems offers tremendous opportunity to use vast quantities of data to advance both health care systems as well as individual-level patient health care,” says Aline Talhouk, an author of the report who is an assistant professor in the Faculty of Medicine at the University of British Columbia and principal investigator at British Columbia’s Gynecological Cancer Research Program (OVCARE). “While federated learning in machine learning systems can be extremely useful in seizing this opportunity, it will be crucial that policymakers work with AI scientists and researchers on the ground to devise ways to address challenges around ethics, privacy and data governance, as well as security.”
The policy brief walks policy-makers through eight technical and ethical-socio-legal challenges to building federated learning projects, along with strategies to address these risks, and policy options for public engagement and consent.
“To date, much of the focus of public discussion has been on protecting data,” says Tania Bubela, Dean of the Faculty of Health Sciences at Simon Fraser University who is also a report author. “We argue that the time has come for meaningful public deliberations about the appropriate balance between privacy risks and the harms of not using data for health research and innovation. We know that patients and their families want to see their health data used to improve care. Federated learning, if deployed appropriately and with carefully thought-out precautions, can be a means to achieving these advances.”
Towards a Proportionate and Risk-Based Approach to Federated Data Access in Canada was published today by CIFAR. Co-authors of the report are: Tania Bubela, Dean of the Faculty of Health Sciences at Simon Fraser University; Ivan Beschastnikh, Associate Professor at the University of British Columbia; Regiane Garcia, Research Associate at Simon Fraser University; and Aline Talhouk, Assistant Professor and Michael Smith Health Research BC Scholar at the University of British Columbia’s Faculty of Medicine.
For more information, contact:
Program Manager, AI & Society, CIFAR
About CIFAR AI Insights
CIFAR AI Insights is a series of policy briefs inviting cross-disciplinary experts to author accessible policy briefs that discuss the practical societal and political implications of AI and emerging technologies. They are designed to develop Canada’s thought leadership on issues of importance to policy-makers, researchers, regulators and others seeking to engage with and address the societal impacts of AI.
About the Pan-Canadian AI Strategy at CIFAR
The Pan-Canadian Artificial Intelligence Strategy at CIFAR drives cutting-edge research, trains the next generation of diverse AI leaders, and fosters cross-sectoral collaboration for innovation, commercialization and responsible AI adoption. Our three National AI Institutes – Amii in Edmonton, Mila in Montréal, and the Vector Institute in Toronto – are the vibrant central hubs of Canada’s thriving AI ecosystem. Funded by the Government of Canada, we’re building a dynamic, representative, and rich community of world-leading researchers who are creating transformative, responsible AI solutions for people and the planet.
CIFAR is a registered charitable organization supported by the governments of Canada, Alberta and Quebec, as well as foundations, individuals, corporations and Canadian and international partner organizations.