Assistant Professor Shion Guha has co-authored a new report calling for action to improve healthcare data sharing. Focused on actionable recommendations to improve healthcare, the report identifies key strategies for establishing trust, promoting equity and navigating the evolving regulatory landscape of AI in healthcare.
The report, which was released in February, grew out of a workshop held by the Computing Research Association’s Industry committee last year. It brought together 35 representatives from industry, academia, and government to discuss current opportunities and risks associated with health data sharing, explore associated challenges of data ownership, access, and control, and identify technological innovations that can overcome these barriers
The report emphasizes that continued engagement is crucial given the evolving landscape of healthcare AI policies and regulations. It puts forward several key takeaways and recommendations for a path forward in the healthcare sector:
- Establishing trust as foundational for encouraging data sharing, particularly with marginalized populations, by partnering with trusted local organizations (e.g., faith-based groups, libraries, and community centers). This also means acknowledging and addressing historical distrust, misinformation, and security concerns as well as enhancing physician-patient communication.
- Promoting data access, standardization, and quality to enable patients to easily access and share their health information, reducing errors and delays. Standardization of data and interoperability of data systems across healthcare and community organizations is crucial for tracking upstream and downstream effects.
- Addressing ethical considerations related to data equity and AI system design, including human oversight in AI systems. The report advocates for ethically designed AI systems that include “human-in-the-loop” oversight, capturing individual values and daily variations rather than simply automating tasks.
- Navigating the regulatory landscape by aligning data practices with standards and promoting a common vocabulary, as well as creating feedback mechanisms for emerging AI policies.
Health informatics is a key research area at the Faculty of Information. Guha’s work demonstrates the Faculty’s commitment to addressing critical issues in data governance, ethics, and policy within the rapidly evolving healthcare landscape.
Featured News
MMSt students’ exhibit on sound in the movies launches at TIFF
For their capstone project, Museum Studies students Liz Sullivan, Lana Zagorac and Devin de Silva collaborated with TIFF’s Film Reference Library to explore how sound effects, music and even silence shape audiences’ experience and linger long after the credits roll. The resulting exhibition, From Silence to Sound: Tuning the Auditory Experience, features objects from the the library’s Christopher […]
In Memoriam: Professor Twyla Gibson (1954 – 2024)
Friends and Colleagues remember Twyla Gibson as an interdisciplinary scholar whose information interests ranged from Plato to Marshall McLuhan While she came to academia later in life than usual, Twyla Gibson, left a big imprint, including at the Faculty of Information, where she held an assistant professor appointment from 2007 to 2011. Gibson was […]
BI student wins leadership award after finding her niche at Faculty of Information
It was love at first course for Sehar Bajwa when she switched to the Bachelor of Information (BI) program from engineering where she had completed her first two years of study at U of T. After two years in a purely technical environment, she realized it was stifling her creativity. “I looked at myself in […]
Activating Indigenous photographic history at Deyohahá:ge:
As Indigenous histories have been systematically erased, many First Nations’ analog records remain uncatalogued and inaccessible. These physical artifacts, including important films and photographs, are particularly at-risk for degradation and loss. Given that archives hold the evidence of history and that the histories we know are the histories available to us, the Faculty of Information’s GLAM […]