Pakistani Classrooms: A teacher in a small classroom in the Malir district of Karachi opens up her phone, writes a few lines in an application, and requests assistance in designing a science lesson. She can create an interactive story of a raindrop named Rani containing a brief Urdu poem in a few seconds. She has students around her, many of whom do not have printed textbooks and are wonder-struck.
Such a scene, which is not common but provides clues of a silent revolution, is that Artificial Intelligence (AI) is infiltrating Pakistani classrooms. Education is being pushed into the digital future with elite schools experimenting with adaptive learning platforms and government pilots, and AI teaching assistants.
However, behind the buzz, there is a more difficult question: would it be possible to integrate AI in Pakistan without further entrenching the very inequalities that are already the hallmark of the Pakistani education system?
The Future of AI in Pakistani Classrooms
A Tale of Two Systems
The classrooms in Pakistan reflect the country’s social differences. On one extreme, we have elite private schools and city universities that have high-speed internet and digital laboratories. Also at the other end are the thousands of government schools that are under-resourced and struggle to sustain even the basic infrastructure.
The Pakistan Bureau of Statistics (2023) estimates that approximately 36 per cent of households are equipped with internet access, which reduces to under 20 per cent in villages and towns. Computer labs and adequate electricity are still not available in many public schools. Therefore, AI-based applications can tailor the lessons to a student based in Lahore, but such a concept will not be familiar to the majority of children in Balochistan or the interior of Sindh.
Moreover, an education report by UNESCO-Pakistan (2024) cautioned that digital technology has the potential to increase inequality in case there is unequal access, training, and language. This is the gist of the issue: AI is supposed to provide personalised learning to already digitally connected people.
The Promise and the Perils of Personalisation
AI is being touted as a disruptor to education across the world. Learning gaps may be detected through algorithms, and customised content may be provided, as well as teachers can receive real-time data on student progress.
These potential findings reflect some local experiments in Pakistan. In the Punjab Information Technology Board, the Taleem Ghar project uses AI analytics to monitor which lessons on TV students are paying attention to. In Khyber Pakhtunkhwa, the Durshal EdTech Lab has started educating teachers in classroom management tools based on AI. Universities such as LUMS and NUST are now employing digital research support and AI-powered plagiarism detection.
Such attempts are an indication of an experimental nature. However, without serious consideration, AI might not decrease but enlarge the learning gap. Elite institutions will become quicker, teaching AI literacy courses as part of their curricula, and the public schools will fall behind, bringing a new hierarchy under a new name.
Teacher Preparedness: The Fragile Backbone
In case AI is to be implemented in classrooms, it is teachers, not algorithms, who should take the lead. However, this is one of the biggest bottlenecks of Pakistan: the absence of training of teachers and confidence.
Syed Hassan Raza and Azib Farooq (2025) surveyed 125 teachers around the country and found that most of them were willing to use AI technologies, but few had any formal preparation to use them. Many people informally utilised AI to create lesson plans or quizzes but were not supported by the institution.
When teachers experiment independently, the issue of AI becoming a novelty instead of a pedagogical tool is likely to arise. Worse, it may enhance anxiety that technology is there to displace them.
Education technology researcher Dr Nadia Khan at Quaid-i-Azam University says that teachers are being instructed to incorporate AI, but they have no idea what AI means or how it functions. They require training emphasising morality, imagination, and regional setting — not the utilisation of software.
This gap may be filled by provincial courses in the form of the provincial AI Master Trainer programmes — brief workshops within the existing teacher education structures. Digital reforms will not become systemic until they are so invested.
Curriculum without Context
AI has been developed with flexibility; the same cannot be said of the curriculum in Pakistan. Strict syllabuses, old-fashioned textbooks, and teacher-centred test-based instruction do not leave much room to ask questions or to be creative — the two aspects that AI might develop.
In a report by the EdTech Hub (2022), provincial governments are said to focus on the digitalisation of classrooms, neglecting curriculum redesign. Consequently, technology will be a mask for rote learning.
It is difficult to imagine that in a classroom where students are still memorising textbook definitions, AI can be used to draft an essay or to interpret data. The future of AI, which has the potential to make one curious, think critically, and become a multilingual innovator, is in conflict with the currently standardised system.
Pakistan would need to reform its definition of learning to make AI meaningful. It implies that we should integrate computational thinking, ethics, and digital literacy in the national curriculum, not as elective courses but as interdisciplinary skills.
The Ethical Frontier
Ethics cannot be left out of the debate on AI in education. Algorithms are not neutral; they are taught through prejudiced information.
Pakistani students who use Urdu-English code-switching or local dialect when speaking English can be mistakenly called to order by AI-based essay graders or chatbots. Which have often been trained on Western data. Applications that can be used to learn languages may implicitly favour standardised English at the expense of local dialects and expressions.
Another threat is surveillance. Other universities have experimented with proctoring software based on AI during online exams. Such systems monitor the movements of faces and the background noise to alert about suspicious behaviour; however, research in other countries has concluded that the systems tend to mislabel students with darker skin complexions or erratic internet connections.
Lack of a powerful data protection law — The Personal Data Protection Bill (2023) is still under review in Pakistan, and student data obtained in AI systems may be abused or sold.
Ethical principles should thus be embedded into all AI pilots: disclosure of the use of data, the right of people not to engage in AI, and the responsibility of algorithmic discrimination. AI will bring the level of learning to a new height, as long as it does not betray trust.
Lessons Learnt from Pilot Projects.
Not all is grim. The positive outcomes of small and local projects demonstrate that AI could be democratised in case it is context-oriented.
The Malir district of Karachi has a pilot (2024) which trained twenty teachers of the government schools to use generative AI in lesson planning. It took place through WhatsApp, as well as in brief workshops, and the change was quite dramatic: in three months, nearly all teachers began to use AI tools at least once a week, and quite a number of them stated that they felt less overburdened with work and that students were more engaged.
The takeaway? The AI does not need costly infrastructure, only the available tools, the training needed, and a mentor. However, the state should fund such endeavours and support them with political will.
The Inequality Dilemma
The 2023 Pakistan Human Capital Review conducted by the World Bank estimates that educational inequality costs about 2% of the GDP each year. AI may address or increase such a loss.
Assuming that the integration of AI follows the existing trend, i.e., the introduction of AI in exclusive schools and subsequently in the general sector. The risk lies in the fact that it may become a divide, known as the ‘AI Divide’ by the researchers, a disconnect between the people who use technologies and those who just look at them.
The issue is to ensure that innovation is inclusive. It entails making AI accessible to schools in the general population, incorporating affordable technology in their native tongue, and creating solutions that can operate without the internet or low-end smartphones.
A Way Forward
Pakistan requires more than the enthusiasm of pilots to balance innovation and justice; it requires strategy. First and foremost are the Public-private partnerships to expand broadband and low-cost devices to underserved schools. Moreover, educate teachers as co-creators, but not as users, of AI-based learning. Additionally, it is necessary to ensure transparency and protect student data under the new AI and digital policy. This is a window of policy that will be offered by the next National AI Policy (2024). Provided it is done with an educational consideration, rather than merely an industrial one, it might establish the basis of a more equitable digital shift.