‘AI is a moving target, and these survey results suggest that schools across Canada have thus far failed to get ahead of it’
HR professionals bracing for a talent shortage of critical thinkers may find the problem is being compounded in Canadian classrooms, as a lot of teachers are not getting the necessary training to detect or manage student use of artificial intelligence (AI) from their schools.
Overall, 64.7% of grades 6–12 teachers have not been provided training or tools to identify when students are using AI to complete their work, and 63.6% say their school or board has not given them guidance on how to instruct students to use AI appropriately, reports the Fraser Institute.
"Undoubtedly, Canadian teachers are dealing with the impacts of AI use professionally and by their students — yet it's clear from these survey results that teachers are not being supported by policies and training in this rapidly evolving space," said Paige MacPherson, senior fellow at the Fraser Institute and author of the report.
A separate Meridian Credit Union survey found that seven in 10 Canadians believe the country is still not doing enough to prepare young adults for work, with one in six saying young adults lack the skills needed to succeed in the job market.
Policy gaps widespread at schools
Only 34.8% of teachers polled say their school has a policy on AI use for staff, and 42.3% say their school has a policy on student AI use, according to the Fraser Institute survey of 756 teachers across Canada.
Just over one third — 35.4% — say their school has no AI policies at all for either staff or students.
"AI constitutes a potentially significant shift in education, reshaping how teachers craft lessons and more concerningly, how many students are actually doing their homework," MacPherson says.
A 2025 KPMG public opinion survey found that 73% of Canadian post-secondary students used generative AI for their schoolwork, and 48% reported their critical thinking skills had suffered as a result. MacPherson's report noted similar trends are likely emerging in upper elementary, middle and high schools.
Also, overreliance on AI tools is making employees doubt their own skills – to the point of feeling less intelligent because of the technology, according to another report.
Schools 'failed to get ahead' of AI
Training rates vary considerably by province. Teachers in Atlantic Canada were most likely to report receiving AI training for in-class use at 61.1%, followed by Alberta at 56.0% and Ontario at 55.4%, reports the Fraser Institute conducted between November 2025 and January 2026. In Manitoba and Saskatchewan, only 24.3% of teachers surveyed reported receiving such training.
Independent school teachers fared better than public school counterparts across nearly every measure. Some 46.6% of independent school teachers said they had received training on instructing students in reasonable AI use, versus 29.2% of public school teachers.
The findings land against a worsening labour shortage. Ontario alone will need about one million university graduates between 2026 and 2035 to fill projected labour shortfalls — about 100,488 per year — according to a Stokes Economics forecast prepared for the Council of Ontario Universities. STEM occupations represent the largest projected demand at 212,980 openings.
MacPherson concludes: “Newer teachers were much more likely than those teaching for 10 years or more, to indicate that school or school board AI policies were in place, and that any type of AI training was provided. This suggests a discrepancy in policy awareness, or perhaps a generational difference in openness to AI policies and training.
“These findings raise serious questions for schools teaching grades 6 to 12, which includes late elementary, middle, and high schools. AI is a moving target, and these survey results suggest that schools across Canada have thus far failed to get ahead of it. “
AI can also cause mental health troubles for the people that use them, according to one expert. Among the challenges are “cognitive overload, managing the AI outputs as well as verifying the accuracy,” alongside concerns about skill erosion and dependency, said Stephanie McNamara, a researcher at the University of Florida.
