The Ambitious Vision of Pakistan’s National AI Policy 2025
Last July, Pakistan’s federal cabinet approved the National AI Policy 2025 with considerable fanfare. The goals were ambitious: a million AI-trained professionals by 2030, AI concepts embedded from early STEM education upward, and universities repositioned as engines of an AI-driven knowledge economy. A few months earlier, the Higher Education Commission had already revised its Computer Science curriculum to make AI a core component of undergraduate study. On paper, Pakistan is moving. Inside the classroom, the picture is rather different, and I say that as someone who sits in those classrooms.
The Systemic Failure in Faculty Training
The uncomfortable truth is this: HEC mandated AI integration into the curriculum without any corresponding programme to train the faculty who are supposed to deliver it. I have watched senior professors, respected in their fields, open a browser and paste a research question into ChatGPT, then tidy up the output and present it as analysis. I am not making a moral judgement about those individuals; I am pointing out a systemic failure. We asked people to teach something they were never taught, gave them no tools to do it responsibly, and then wondered why the results look the way they do.
The Student Dilemma and the Rise of AI-Generated Assignments
The student side of this is equally troubling, and arguably more consequential. AI-generated assignments are now routine. But here is the part that does not get discussed enough: the humanising software has caught up with the detection software, and it has won. Tools that rewrite AI-generated text to mimic human patterns are freely available, widely used, and largely invisible to the plagiarism checkers most Pakistani universities rely on. A student can generate an essay, run it through a humaniser, and submit it without a single flag. The system has no answer for this, not yet. So, we have arrived at a strange equilibrium: everyone knows what is happening, nobody can prove it, and the degree keeps getting awarded.
The Erosion of Academic Integrity and Critical Thinking Skills
What is lost in all of this is not just academic integrity. It is the actual skill that a degree is supposed to certify. Critical thinking, the ability to take a concept from a lecture and connect it to something real, to interrogate an argument rather than just reproduce one, is precisely the capacity that gets bypassed when AI does the thinking. And this is the skill that the global job market is increasingly demanding, because it is the one thing AI does not do well. We are producing graduates who are fluent in retrieval and helpless at reasoning.
Moving Beyond Banning AI: A Path Forward
I want to be careful here, because the answer is not to ban AI use or to pretend the tools do not exist. That ship has sailed, and performative bans only push the problem underground. The answer is to be honest about what ethical AI use actually looks like and to build that literacy deliberately – starting with the faculty. HEC’s revised curriculum lists AI as a course outcome. It says nothing about how a professor who has never used these tools in a disciplined, reflective way is supposed to model that use for students. That gap needs to close, and it needs to close before the next intake arrives.
Practical Steps for Institutional Change
What would that look like in practice? Faculty development programmes, not optional seminars but structured, assessed training, on both the capabilities and the limits of AI tools. Revised assessment formats that make AI-generated submissions structurally harder to pass off, not just easier to detect. And an honest institutional conversation about the humanising software problem, because pretending our current detection tools are adequate is its own kind of intellectual dishonesty.
The Gap Between Policy and Reality
Pakistan’s AI ambitions are not unreasonable. The policy exists. The curriculum revision exists. The intent, at the level of government, appears genuine. But intent without implementation is just paperwork. And right now, the implementation is being left to students and faculty to improvise on their own, with predictable results. If we are serious about the million trained professionals target, we need to start with the people already in the room.
