OTTAWA — Richard Sutton thinks a lot about thinking. The University of Alberta professor helped instigate the development of modern reinforcement learning, a school of artificial intelligence techniques that taught a computer program the game of Go and is letting the little oil-sands town of Drayton Valley, Alta., wring efficiencies from its water-treatment plant. “The big prize, as far as I’m concerned,” he told The Logic in an interview late last year, “is really understanding how people work and making something of comparable abilities to automate lots of the dull, boring, dirty or dangerous activities that can be taken over by machines.”
The federal government is preparing to spend hundreds of millions to expand its national AI strategy, and replicate it for quantum. Proponents say the first incarnation of the initiative, focused on recruiting and retaining top researchers, has helped fuel the growth of the AI ecosystem, but some scale-up executives are concerned a lack of focus on commercialization risks squandering Canada’s AI advantage.