I've been told my numerous people who grew up there that the winters in Edmonton were brutal. Despite Toronto's reputation being a Canadian city the winters are very mild by comparison and Vancouver, of course, is even easier.
Having lived in Montreal on the other hand I was surprised at the impact a lengthier far snowier winter can have on a city if you didn't grow up with it, after being familiar with Toronto/southern Ontario I found the winters in MTL hard to deal with. A winter-heavy city really does create two very different cities, the summers in Montreal are really amazing though, half the year it's the best city in Canada.
Keeping people inside for winter may be a non-negative in academic research but it's probably the most serious implication to consider lifestyle-wise when moving to Edmonton. Not to mention the lack of access to events (both culturally and industry-wise).
I'm not sure I could be optimistic for a Canadian city for technology outside of Toronto, MTL, and Vancouver, besides the current available talent pool which can create at least a decade-long attraction with someone exceptional like Rich Sutton. Similar to sports teams talent attracting free agents. Which is what we're talking about in this case, with advanced research in AI.
The Manhattan project's success was largely a result of the talent they attracted early on, which blossomed into even more talent as a result, and it was located in the middle of nowhere in an isolated New Mexico town.
City-wise it might not be attractive but talent is what matters at the end of the day. The only risk then is that it can maintain an active research community, aka consistently attracting talent. I'd be measuring the output of the universities with highest priority in this case more than anything.