AI is a rather meaningless term, in reality it means that a data set is selected by a select group that will in all probablity not recognize their bias.  After several years, someone will notice that the AI results match the Boots on the Ground bias, and then blame the programmers rather than the Boots.

I would ask that ALL AI projects are reviewed by panels that are both technically qualified and also reflective of the makeup of the city, and include proportional representation of women, BIPOC, religious, disabled, and such other communities as feel that inadequate development of AI puts them at risk.

It can be as simple as Cadillacs don't get spot checked, or people working overnight are unfairly targeted, but there are people available who can develop systems for Toronto that are equitable rather than use American systems known to be biased.