By: Erin Vollick
9 Oct, 2023
Sheila McIlraith, Vector Institute: Building safe AI for the generations to come
October 9, 2023
Sheila McIlraith, Canada CIFAR AI Chair, Vector Institute
Sheila McIlraith, a Canada CIFAR AI Chair at the Vector Institute, wants to make AI systems human compatible – and safe. A leader in the field of AI from its early days, her recent research focuses on the ways language can support AI automated planning and decision-making.
“The impact of CIFAR and the Pan-Canadian AI Strategy has been significant, attracting a critical mass of exceptional AI researchers, and supporting a training environment for students that is second to none — all within a vibrant Toronto tech community, and all in Canada.”
Sheila McIlraith, Canada CIFAR AI Chair, Vector Institute
McIlraith and her students, who she enthusiastically describes as “awesome,” were the first to successfully demonstrate that by exploiting the syntax and semantics of language, AI agents learned more quickly. For this work, they earned the 2023 IJCAI-JAIR Best Paper Prize. In 2022, McIlraith received the International Conference on Automated Planning and Scheduling Influential Paper Award for a high-impact 2012 paper that addressed the fundamental AI challenge of generating plans to achieve goals in complex settings.
Passionate about addressing safety issues that come with human-compatible decision making, McIlraith, who is also an associate director at the Schwartz Reisman Institute for Technology and Society (SRI) at the University of Toronto, devotes considerable research to this topic. She additionally worked with computer science, SRI, and philosophy colleagues to develop a first-in-Canada embedded ethics program for undergraduate curricula that launched in 2020. The program aims to empower the next generation of scientists, educators, and technology developers with the knowledge, skills, and incentives to incorporate ethical considerations in the study and design of technology systems. Next year, approximately 7,000 students across the University of Toronto will learn through the program, and the university will host an event to support other Canadian universities in developing similar programs.
“The impact of CIFAR and the Pan-Canadian AI Strategy has been significant, attracting a critical mass of exceptional AI researchers, and supporting a training environment for students that is second to none — all within a vibrant Toronto tech community, and all in Canada,” says McIlraith.