“Wild West”: Teachers and experts seek unified policies for AI in classrooms

Canadian schools are trying to decide how artificial intelligence should enter the classroom, but educators say the rules vary too widely from one jurisdiction to another. The result, one teachers' leader said, resembles a "Wild West" approach to a technology developing faster than school policy.

“

Wild West

Jason Schilling, president of the Alberta Teachers' Association, said schools, boards and governments had issued little consistent guidance. Much of the responsibility has fallen to individual school districts, leaving teachers to work out how students should use AI while preserving critical thinking and intellectual honesty.

Michelle Carlos of Winnipeg's Pembina Trails School Division said the lack of common rules creates inconsistency across the education system. Her division developed its own guidelines about two years ago after waiting for government direction that did not arrive. It has not enabled AI tools on student networks, although teachers can use them. Teachers are encouraged to explain when AI may support an assignment and when it should not be used.

“

There’s a lot of concerns around (AI), but there’s not a lot of policy that you’re seeing at the jurisdictional level, at school boards or at government

Carlos described AI as a "thought partner" rather than a "work doer." She supports national or provincial frameworks, saying policy development takes time and resources while schools are also being urged to innovate and manage the risks.

“

It can be a ‘thought partner’ but not a ‘work doer’

Manitoba Education Minister Tracy Schmidt said her department is developing guiding principles to help schools make "thoughtful, ethical and safe decisions." A Leger survey for the Fraser Institute found that 49% of 756 teachers in grades 6 to 12 had received training from their school or board on using AI in lessons. About 63% said they had received no training on teaching students to use AI reasonably. The online survey ran from November 2025 to January 2026, and no margin of error was assigned because respondents were not selected randomly.

“

thoughtful, ethical and safe decisions