Pakistan's universities are not facing a teacher shortage. They are facing a relevance shortage. In faculty corridors, the question is still whispered with unease: will artificial intelligence replace professors? In the offices of industry, at the HEC, and in Board of Studies meetings where curricula are approved, a far more consequential question must now be asked: which professors and which institutions are using AI to teach, assess, and mentor in ways that were impossible five years ago? That is the divide that will define the next decade of higher education in this country. Institutions that answer it honestly will flourish. Those that pretend it does not exist will quietly become degree-printing houses with shrinking credibility.
For most of the last century, the contract between a university and its students was elegantly simple. The professor was the gatekeeper of knowledge delivering lectures, setting papers, and passing judgment on answers. That contract has now been broken, irreversibly. AI can summarize a three-hundred-page monograph before a student finishes the breakfast, generate multiple explanations of a single concept, and assess a short-answer quiz with reasonable precision. If we continue to define teaching as the transfer of information, the anxiety surrounding AI is understandable. But if we redefine teaching as mentorship, ethical modeling, and the cultivation of judgment, then AI becomes the most powerful pedagogical assistant a professor has ever been offered. The problem is that most of our systems and incentives remain stubbornly built for the old definition.
This mismatch is already producing two very different kinds of classrooms. In the first, AI is officially banned. A circular is issued by QEC or the academic departments, shared on the university's Facebook page, and forgotten before the week is out. Students use ChatGPT and Claude discreetly to complete assignments while teachers spend weekends marking hundreds of near-identical, intellectually shallow assignments. Learning outcomes stagnate and the prohibition achieves nothing beyond institutional self-deception. In the second kind of classroom, AI is treated as infrastructure just like a library or the laboratory. Students engage with it under explicit conditions: they submit prompts, document iterations, and defend their final work before a faculty panel. Professors use AI to give every student personalized feedback within forty-eight hours and reclaim time for the mentorship only a human being can provide. Now, consider two graduates, one from each of these classrooms, presenting themselves to an employer in Islamabad, Dubai, or London. The choice will not be difficult. The real fault line in Pakistani higher education is no longer between public and private institutions. It is between AI-literate institutions and AI-illiterate ones.
What passes for AI capacity-building on many campuses is an exercise in institutional optics: a one-day seminar, a printed banner, a few social media posts, and no measurable change in what happens inside a classroom the following Monday.
The appropriate response to this reality cannot be a faculty workshop followed by a group photograph. What passes for AI capacity-building on many campuses is, an exercise in institutional optics: a one-day seminar, a printed banner, a few social media posts, and no measurable change in what happens inside a classroom the following Monday. Real capacity-building is slower and requires sustained academic engagement. It begins by asking an uncomfortable question about every course outline currently in use: if a competent student can fulfill this assignment with a single AI prompt, what exactly is being assessed? If the honest answer is memorisation or the reproduction of lecture notes, then the assignment was already pedagogically inadequate before AI arrived. The technology has simply made that inadequacy impossible to ignore.
The most consequential site of reform is the Board of Studies. Consider one concrete illustration. In a conventional BBA International Business course, students are asked to write a report on a multinational corporation and are evaluated through a conventional final term exam. In an AI-integrated version, the assessment becomes: use AI to identify three viable export markets for surgical instruments from Sialkot, submit your prompts alongside the AI output, validate those outputs against real trade data, and defend your recommendations. The substantive content is comparable. What changes is the cognitive domain being assessed from recall to judgment and the teacher role shifts from information provider to mentor and analytical coach. This is not a theoretical proposal. It is a course outline revision any Board of Studies could approve this semester. The same principle extends across disciplines. The curriculum must consistently demand that students operate one level of reasoning above what the machine can independently produce.
This necessarily transforms how we assess learning. Three-hour papers composed of definitional questions effectively assess little more than a student's willingness to refrain from consulting AI in an unsupervised room. A more coherent framework requires process-based assessment, where students submit their full working prompts, rejected sources, and a reflection on where AI misled them alongside the final product. It requires defense-based assessment, where students account in their own words for every choice embedded in their submitted work. AI cannot sit a viva on a student's behalf. Medical and law schools have understood this for decades. The remaining challenge is to mandate this across disciplines through curriculum committees rather than leaving it to individual faculty initiative.
None of this is achievable without faculty who are genuinely equipped rather than merely photographed at a seminar. HEC and university leadership must treat AI training as academic infrastructure and fund it accordingly not as an event, but as a structured twelve-week programme in which faculty redesign actual courses, review each other's work, and are evaluated on whether student learning outcomes noticeably improve. Promotion committees should ask candidates to show how they have used technology to improve teaching quality. A professor who redesigns a course to deliver personalized feedback to four hundred students deserves institutional recognition no less than a colleague who has published journal articles. At present, our incentive structures reward only the latter. Boards of Studies are precisely where that must begin to change.
Artificial intelligence will not close Pakistan's universities. But an institution that has genuinely redesigned its curriculum and assessment will, over time, make an AI-ignoring peer irrelevant. Students will vote with their enrolment decisions. And employers will vote with their hiring. The academic profession is not disappearing rather it is transforming, steadily and irreversibly, toward educators who have learned to use these tools with intellectual rigour and toward institutions that had the courage to act before circumstances forced their hand. The choice before every Vice-Chancellor, Dean, and Board of Studies member is therefore straightforward. We can treat AI as a threat and respond with bans and banners. Or we can treat it as a pedagogical utility and do the slow, genuinely important work of redesigning what we teach and how we measure learning. That work will not trend on social media. But it is the only work that is truly appropriate with what we actually need.