The race to reimagine higher education

“Universities are among the most durable institutions human beings have ever created. While a scholar from the Middle Ages might have found parts of the modern campus bewildering, they would still recognize the basic form: experts at the front of rooms, students organized into courses, knowledge divided into disciplines, credentials awarded after examinations. For all the technological change around them, universities have remained remarkably stable because their core product has always depended on something difficult to capture and mechanise: expert tacit knowledge. For that same reason, they are now about to be transformed. The real significance of artificial intelligence is not that it can write essays, summarize documents, or answer emails. It is that, for the first time in history, machines can capture tacit knowledge: the practical, experience-based know-how that experts possess but cannot fully explain. It is this tacit knowledge that has made doctors, lawyers, professors, and other experts so valuable in the current economy. Machines could not do what we ourselves could not write down. Machine learning changed that. Computers no longer require codified instructions to automate a task. Instead, machine learning observes contexts, actions and outcomes, and infers patterns. A model trained on clinical interactions can begin to perform diagnosis. A model trained on hundreds of thousands of customer-service conversations can capture some of what sets the best agents apart from the rest. That tacit knowledge can then be used to automate customer service, or to augment workers who do not possess the same expertise. Recent work shows this clearly: an AI conversational assistant increased productivity especially among less experienced customer-service representatives by helping them respond more like highly skilled ones. Education is one of the most tacit-knowledge-intensive sectors in the economy. Great teachers do not merely transmit information. They sense confusion before students can articulate it. They choose examples that land. They slow down, speed up, reframe, encourage, challenge and diagnose. Professors do not just impart knowledge from textbooks, they contextualise it using their specialized tacit expertise. A good supervisor knows when a graduate student needs freedom to explore and when they need structure, and passes on the craft of doing research. A good admissions officer, fundraiser, co-op adviser or department chair often relies on judgment learned through years of experience. Activities at the heart of education like teaching, advising and research, have been difficult to automate or scale without sacrificing quality. As a result, higher education has long faced Baumol’s cost disease: productivity growth in more technologically dynamic sectors has pushed wages upward across the economy, while universities have continued to require large amounts of expert human time, creating persistent upward pressure on the real cost of high-quality education. AI breaks that pattern. Or, more precisely, it creates the first credible opportunity to do so. For the first time, we have a technology that augments or automates most of the activities that are core to universities. This is why universities are so exposed. Rethinking higher education in AI-integrated world This does not mean that professors disappear, that universities become apps, or that campuses empty out. Instead, activities once limited by scarce expert time become abundant. Tutoring, advising, feedback, translation, explanation, research synthesis and customized learning can be delivered at a scale that was previously unimaginable. The question is not whether universities will be transformed. The question is whether Canadian universities will lead that transformation or have it imposed on them by technology companies, global competitors and impatient students. In some of my recent work, I developed the 3R framework — Replace, Reimagine and Recombine — to lay out how general-purpose technologies reshape organizations and economies. The same framework applies to universities. The first phase is Replace: using AI to perform existing tasks faster, cheaper, better or at greater volume, by replacing the old technology. This is where every university should begin. The evidence is already compelling. In a controlled experiment, ChatGPT improved productivity on professional writing tasks, reducing time and improving quality. In another experiment, GitHub Copilot users completed a programming task 127 per cent faster. A study with Boston Consulting Group found that consultants using GPT-4 completed more tasks, worked faster and produced higher-quality outputs. For universities, the Replace opportunities are everywhere. In teaching, AI can help generate course content, including outlines, lecture notes, slides, examples, simulations, assignments, exams, and rubrics. It can also help with grading, providing feedback to students, and drafting course communications. In research, it can conduct literature reviews, propose novel research questions, collect, clean and analyse data, develop software tools, prove theorems, help write papers and assist in grant applications. In operations, AI can improve hiring, marketing, fundraising, reporting, accounting, procurement, and scheduling, and increase student and staff access to information. But Replace should not be confused with strategy. It is merely the first step. The second phase is Reimagine. This is where the real disruption occurs. Electricity did not transform manufacturing simply because factories replaced steam engines with electric motors; the true gains came when factories were redesigned around the possibilities of distributed power. Likewise, computers did not transform business simply by replacing typewriters and calculators. The larger gains came when firms rethought supply chains, customer relationships, global operations and organizational structures. Business models built around the internet beat legacy business models that merely used the internet. Just ask Barnes & Noble or Blockbuster. The same logic now applies to higher education. The clearest Reimagine opportunity is teaching. For decades, Benjamin Bloom’s “two-sigma problem” framed one of education’s central dilemmas: one-on-one tutoring combined with mastery learning could dramatically outperform conventional classroom instruction, but it was too expensive to scale. AI changes the economics. Early evidence is already emerging. A 2025 randomized controlled trial in a Harvard undergraduate physics class found that students learning through an AI tutor learned significantly more in less time than students in an active-learning class, and they reported feeling more engaged and motivated. The university course of the future will not be a recorded lecture with a chatbot add-on. It will be personalised, mastery-based and adaptive. Students will move at different speeds. The AI tutor will know what they have mastered, what they misunderstand, what teaching approaches work for them, what examples resonate and when they are ready to advance. The professor will become less a broadcaster of generic content and more a designer of learning environments, a mentor, a provocateur, a disciplinary guide and a guardian of standards. This will require us to rethink assessment. If AI can write a competent essay or solve a problem set, these cannot remain proofs of learning. We will need more oral exams, live problem-solving, team-based simulations, project portfolios, and lab demonstrations, each followed by short defences where students explain their reasoning, choices, evidence and use of AI. The good news is that AI itself will make alternative modes of assessment, like oral exams, more scalable. AI tutors will blur the line between teaching and assessment because evaluation and feedback can happen continuously. Research will also be transformed. AI can help spot gaps across fields, identify contradictions in literatures, propose hypotheses, generate candidate materials or compounds, design experiments and simulate alternative explanations. The research enterprise becomes less like a lone scholar searching a library and more like a human-AI discovery system. The universities that lead will build shared research platforms, trusted data infrastructure, secure model environments and AI research assistants. Just as “dark software factories” are beginning to emerge, in which AI agents write, test and deploy code with minimal human supervision, some university research groups may become dark research factories: labs where AI agents do the bulk of the work and human researchers focus on setting standards, framing questions, evaluating outputs and governing the overall system. Operations offer equally important transformational opportunities. Universities should be building AI-enabled advising systems that monitor student progress in real time, identify risks early and prompt timely human intervention. Admissions can be reimagined from a passive application-processing system into proactive talent discovery, especially for underrepresented students who may never imagine themselves at a particular university. Co-op and employer relations can become more targeted, using AI to identify emerging employers, match students to opportunities and help firms understand how to use student talent in AI-enabled workplaces. Fundraising can become more personalized and strategic, not by replacing human connection, but by freeing staff from administrative burden and giving them better insight. University curricula must change just as profoundly. The labour market will increasingly punish skills that are close substitutes for AI and reward skills that are complements. Writing, coding, editing and background research will not entirely disappear, but their market value will decline. The premium will shift toward problem framing, entrepreneurship, critical thinking, decision-making, taste, judgment, leadership, communication and the ability to work effectively with intelligent tools. Every graduate, in every discipline, must leave university AI-literate. Not because every student will become a technologist, but because every profession will be reshaped by AI. All of this must be governed carefully. Universities have obligations around accessibility, privacy, fairness, academic integrity and public trust. Those responsibilities are real, but they cannot become an excuse for paralysis. A university that tells students and staff not to use AI is like a city banning electricity because early wiring caused fires and electrocutions. The risks were real; the answer was standards, regulations, inspections, and training, not preserving gaslight. The responsible path is institutional adoption with clear guidelines, secure tools, transparent accountability and a culture of experimentation. The third phase is Recombine. AI will combine with robotics, sensors, the Internet of Things, extended reality, synthetic biology, and quantum computing, among others, to create entirely new technologies. It is hard to predict these combinations with confidence. That is precisely why universities matter as technology research centres. We are among the few institutions designed to scan horizons, combine disciplines and explore possibilities before markets know what to do with them. The future of universities will in part depend on whether we can turn campuses into living laboratories for recombination. None of this means the university campus becomes obsolete. Quite the opposite: campus and student life become essential. If content becomes abundant, discussion and debate become crucial. If tutoring becomes scalable, mentorship becomes more important. If AI can generate solutions, the human capacities to choose good questions, take risks, and lead teams become more valuable. The residential experience, clubs, athletics, design teams, student government, labs, entrepreneurship programs and co-op placements are not peripheral to the AI university. They are central to it because they help students develop the leadership, judgement, interpersonal, and other human skills whose value rises in an AI-rich world. The AI moment is a structural break — and an opportunity to build a university that is more personal, ambitious and human than the one around us today. The greatest danger is not that universities move too quickly. It is that they move too slowly, allowing others to define the university of the future on our behalf. Elite global institutions, private platforms, employers, and students will not wait for our governance processes to become comfortable. The AI university is coming. We can build it in a way that strengthens public education, expands access and opportunity, supports research excellence and deepens the human experience of university life. Or we can defend the old model until it is too late. Universities have survived for centuries because they adapt slowly to preserve what matters. But there are moments when preservation requires reinvention. This is one of those moments. The post The race to reimagine higher education appeared first on University Affairs .

Summary generated from the RSS feed of University Affairs CA. All article rights belong to the original publisher. Click through to read the full piece on www.universityaffairs.ca.