This is Part 3 of Harper Lane’s look at how artificial intelligence (AI) is disrupting higher education. If you have not read Parts 1 and 2, here are the links:
In Part 3, Harper describes how institutions of higher learning are attempting to shut AI out or adapt to it by redesigning teaching approaches, curricula and grades. As always, your comments are welcome.
Universities Rewriting the Higher Education Social Contract
How can higher education adapt curricula in the presence of AI? AI has weakened the connection between assignments and learning. It has exposed incentive structures that increasingly reward credentials over understanding.
If you have read Parts 1 and 2, you may begin to understand just how disruptive AI is to existing post-secondary institutions.
Diagnosing the problem was the easy part. The harder question for universities and colleges is what to do about it. Many search for tools that detect AI to prevent its use altogether. Others embrace AI enthusiastically and attempt to integrate it into every aspect of teaching and assessment.
Both approaches risk missing a deeper reality: higher education is unlikely to return to the world that existed before generative AI. If universities want to preserve learning as the central purpose of education, they need to redesign their courses and how they assess a student’s progress.
The Case for Bringing Learning Back Into the Classroom
Many commentators assume that AI-resistant education requires expensive tutorials, Oxford-style discussions, or highly personalized instruction. But there is nothing inherently resource-intensive about assessing learning directly. (If anything, AI is much more resource-intensive.)
Large lecture courses long existed before personal computers. Students completed exams on paper. They wrote essays in blue books. They demonstrated knowledge in controlled environments. Professors evaluated their work personally. Knowledge was assessed without external assistance.
Assessment Without Tech Surveillance
A modern version of this model can work surprisingly well. Imagine a course where lectures remain large. Instead of submitting weeks of take-home assignments, students demonstrate learning through in-class writing, problem-solving, presentations, examinations, and structured discussions.
In this system, homework exists primarily as optional practice rather than graded completion exercises. Study guides, recorded lectures, and digital resources help students prepare for in-class activities.
Such a system would not eliminate technology. Students could use AI extensively while studying. They could use it to generate explanations, explore examples, and practice concepts.
What they could not do is substitute AI performance for their own performance when demonstrating mastery of a subject.
A New Division of Academic Labour
Perhaps the most interesting possibility is organizational rather than technological.
Universities should consider separating knowledge-focused from tool-focused education more explicitly:
A theory-oriented course (“What do you understand?”) would emphasize conceptual understanding, reasoning, and intellectual development. Assessment would focus on what students personally know and can explain.
An application-oriented course (“What can you accomplish with modern tools?”) would embrace AI directly. Students would learn prompt engineering, workflow automation, verification techniques, model evaluation, and human-AI collaboration.
Currently, many courses attempt to answer both questions simultaneously. That approach is increasingly untenable.
A student who uses AI to generate sophisticated business reports may demonstrate excellent technological competence while revealing little about their underlying business knowledge.
Conversely, a student who demonstrates strong analytical skills without using AI may still be unprepared to use it at an employer’s request in the modern workplace.
The Next Decade of Higher Education
Universities are unlikely to choose between fully embracing AI and fully rejecting it. The future will probably be more nuanced.
Some courses will become AI-native environments where students learn alongside intelligent systems. Others will become intentionally AI-limited spaces designed to cultivate foundational knowledge and reasoning.
Assessment will shift away from monitoring process and toward verifying mastery. Homework may become less important. On the other hand, in-class demonstrations of understanding may become more important. Grades may reflect what students know rather than what they can submit.
The bigger question is what happens if universities fail to adapt.
What Happens if Nothing Changes?
For centuries, a degree served as a reasonably reliable signal that its holder had acquired a particular body of knowledge and way of thinking. If employers, graduate schools, and society at large begin to lose confidence in that signal, universities may find themselves facing a crisis of legitimacy far more serious than any debate over plagiarism or academic misconduct.
The institutions that succeed will distinguish between teaching students how to think and reason, and teaching students how to use increasingly powerful tools. Both are valuable goals, but they are no longer the same.
Educating Humans in the Age of Machines
The student who eventually graduates into this new AI-saturated world will depend in part on their ability to work with and alongside AI. But the student will also depend on something AI cannot easily provide: judgment, expertise, creativity, and the ability to recognize when the AI is wrong.
In some respects, curricula may look surprisingly old-fashioned; not because universities reject technology, but because AI forces them to rediscover their original purpose.
In a world where information, summaries, essays, and even analyses can be generated instantly, the most valuable things a university can teach will be how to think and reason, how to formulate questions, how to develop arguments and hypotheses, and how to discern fact from fiction.
