AI is disrupting education. Universities must change, and change fast

Joël Blit is an associate professor of economics at the University of Waterloo, a senior fellow at the Centre for International Governance Innovation, and the co-founder and co-director of the Canadian AI Adoption Initiative.

As students return to campus this fall, they will encounter updated policies governing the use of AI, yet find a university experience that remains largely unchanged. That is a shame. AI is changing the economics of education, and universities must rethink what they teach and how they teach it.

Education may ultimately be the sector most transformed by AI because it relies so heavily on tacit knowledge: Know-how that experts possess but cannot fully write down. Great teachers curate content, choose examples, read the room, adjust the pace and motivate students. Until now, this expertise has resided exclusively in people.

Opinion: The question isn’t whether AI can teach. It’s whether students will let it

Universities have therefore been organized around a basic constraint: The scarcity of expert time. Educators have known for decades that mastery-based, one-to-one tutoring produces much better learning outcomes than conventional classroom teaching. But society has never had the resources to provide that kind of personalized instruction at scale. Instead, professors teach large classes of students through common lectures, readings, assignments and exams. This achieves scale, at the expense of learning.

AI loosens that constraint. AI tutors can diagnose prior knowledge, explain concepts, choose examples, ask leading questions, generate practice questions and provide timely feedback. In a 2025 randomized controlled trial in a Harvard undergraduate physics course, students learning through an AI tutor learned more in less time than those in an instructor-led class. They also reported greater engagement and motivation.

Yet higher education’s response remains largely defensive. Many instructors have banned AI, arguing that it allows students to outsource the thinking, writing and problem-solving that assignments are meant to develop and assess. The concern, while real, belies incrementalist thinking. While AI can indeed undermine current teaching approaches, the opportunity is not to layer AI onto today’s teaching model, but to rethink the model itself.

In the age of artificial intelligence, professors try to make cheating hard again

By leveraging AI, teaching can become mastery-based, personalized and adaptive. Students could work through material until they master it, instead of being left with gaps in knowledge because the term has ended. Content, examples and explanations could be tailored to each student’s needs, and feedback provided continuously. Pace and difficulty could also be adjusted in real time.

The role of the professor would change from broadcaster of knowledge, to guide and mentor; spending more time designing the overall learning experience, leading discussions, posing difficult questions, challenging assumptions and connecting with students. This would constitute a better use of the scarcest resource in education.

Perhaps paradoxically, AI would also make the campus experience more important. As content delivery becomes abundant, seminars, labs, design teams, clubs and co-op placements become more valuable as ways to help students grow and develop skills such as judgment, initiative, communication, teamwork and leadership. Used well, AI can make universities more human, not less.

Beyond how they teach, universities must also rethink the curriculum itself. Tomorrow’s most valuable skills will not be the same as today’s. As AI makes competent writing, coding, summarization and background research more abundant and cheaper, the premium will shift toward the skills that complement it: Problem framing, critical thinking, judgment, leadership and entrepreneurship.

Opinion: We’ve convinced students they shouldn’t sound like themselves when they write. No wonder they use AI

First-year students arriving on campuses this fall will graduate in the 2030s, when AI will be embedded across much of professional life. Every graduate must therefore leave university AI-literate. They must be able to frame a problem, direct an AI system, improve its output, verify its claims and take responsibility for the result. This literacy must be tailored for each discipline, because economists, engineers, nurses and lawyers will use AI differently and face different challenges and risks.

Beyond teaching, universities should use AI to improve student advising and support, accelerate research and improve administration, while setting clear standards for privacy and accountability.

AI will transform higher education. Canadian universities can shape it for the benefit of students and society, or allow technology companies and international competitors to define the sector’s future. This fall, every university must move beyond sanctioning the use of AI and begin reimagining teaching, student support and research. Back-to-school cannot mean back to business-as-usual.