U of T's AI Task Force Releases Recommendations
Source: utoronto.ca
The University of Toronto’s artificial intelligence task force has made public several high-level recommendations. These recommendations are geared toward making the university an “AI-ready” institution, all while upholding its core missions and values.
The Task Force on Artificial Intelligence’s final report seeks to position U of T as a post-secondary leader in carefully using this technology in areas ranging from teaching and research to student services and administration.
The report’s approach focuses on people and includes four broad recommendations: increasing AI knowledge and expertise across the university; creating AI-ready infrastructure, including an “AI Kitchen” for project development; establishing AI response teams for support; and forming an AI adoption advisory table for guidance.
According to Susan McCahan, special adviser to the provost on artificial intelligence and associate vice-president and vice-provost, digital strategies, responding to AI is essential. She believes U of T can lead by demonstrating a value-based and principles-based approach to working with the technology thoughtfully.
The task force emphasized the rapidly evolving nature of AI, including generative AI and large language models like ChatGPT, along with machine learning for pattern discovery and data analysis. The report avoids specific recommendations that could quickly become outdated.
While acknowledging risks like bias, misuse, environmental impact, and undermining explainable decision-making, the report highlights potential benefits. These range from enhancing accessibility and research collaborations to supporting learning, discovery, and innovation.
Working Groups
To develop its recommendations, the task force engaged six working groups. These groups focused on teaching and learning, research, student services, administration, operations, and institutional technology.
Timothy Chan, co-chair of the task force’s research group and U of T’s associate vice-president and vice-provost, strategic initiatives, sees great potential in AI. He believes an open mind is necessary and that AI can greatly benefit the research community if used properly.
The research group suggested using AI tools to streamline research administration, developing protocols for research integrity when using AI tools, enhancing cross-disciplinary collaborations with AI, and providing training for graduate students.
Chan mentioned that AI can significantly speed up literature reviews and data analysis. He helped create an AI model for the milk bank at Mount Sinai Hospital that predicts nutritional composition and generates a daily recipe for pooled breast milk given to hospitalized babies. According to Chan, fat and protein content increased, and variability decreased.
The operations and planning group, chaired by Ron Saporta, chief operating officer, property services and sustainability, experimented with AI tools. The group concluded that AI could improve forecasting, automate repetitive tasks, and enhance outcomes.
Working group members developed a chatbot grounded in U of T’s publicly available institutional procurement policies to test this. Staff inputted procurement questions into the chatbot, verified the response, and sent it back. Saporta stated that the chatbot was correct most of the time and helps staff by handling low-value work.
Other AI Initiatives at U of T
McCahan mentioned that the Navi virtual assistant uses an older form of AI to help students find campus information such as mental health services. A pilot program to help faculty members build AI tutors for their courses is set to expand. Other initiatives include access to a secure version of ChatGPT and the launch of Digital Learning Innovation’s GenAI Literacy Open Educational Resources.
In the spring of 2024, Trevor Young, U of T vice-president and provost, established the task force to conduct consultations, identify risks, and develop principles. This followed the Provost’s Advisory Group on Generative AI, created earlier to assess generative AI tools and create FAQs. Other resources include U of T Libraries’ Generative AI Tools and Copyright Considerations and the School of Graduate Studies’ Guidance on the Appropriate Use of Generative Artificial Intelligence in Graduate Theses.
Young expressed gratitude to the task force and stated that the recommendations will guide U of T in embracing AI and bolstering its role as an AI leader.
The university will now consider how to adopt and implement the report’s recommendations. McCahan believes the report outlines a flexible framework for embracing AI, considering risk, supporting learning, and ensuring U of T maintains its approach. She added that collaboration is essential for making wise choices.