11/21/2022
We've had computers and even for a long time now.
What's different, explained Gillian Hadfield, Director of University of Toronto's Schwartz Reisman Institute for Technology and Society, during her keynote, is that it's evolved to include , combined with mass digitization and huge computing power.
However, it's critical to understand that machine learning is NOT like conventional programming, she said. With conventional programming, the program is written by humans, but with machine learning a computer writes the rules — presenting regulatory challenges.
With machine learning, the machine solves problems and build patterns that we can't see. As a result, Hadfield outlines, there can be a value alignment problem, as the machine might solve problems in a way that we don't want it to.
Another challenge with regulation, Hadfield said, is that sets are massive and constantly evolving. The speed of innovation is very rapid — can regulation keep up?
Hadfield also proposed that AI regulation, as a sector, poses a tremendous business opportunity. Regulatory tech could be developed in Canada, then marketed and sold to other regions around the world — you just have to meet local governance standards.