KHYBER Pakhtunkhwa has introduced an artificial-intelligence teaching model for grades six to twelve. Developed by the Elementary and Secondary Education Foundation, the Global Virtual School promises lessons, assessments, progress tracking and laboratory exercises. Its AI teacher explains the syllabus in Urdu, English and Pashto, answers questions and offers personalised support. It forms part of a push, alongside Parho Khyber Pakhtunkhwa and a proposed AI Education Authority, to place artificial intelligence at the centre of public schooling. Where geography, insecurity and shortages of specialist teachers restrict access, the appeal is understandable.
The initiative deserves neither reflexive celebration nor rejection. Technology can widen access and assist teachers in difficult conditions. A recorded lesson does not fail to arrive because a road is blocked; a virtual laboratory can demonstrate processes that a school without equipment can only describe. Multilingual explanations may help children whose home language differs from the language of instruction. Used carefully, artificial intelligence could become a valuable instrument. The danger begins when an instrument is mistaken for an institution. That distinction will determine whether the programme strengthens public education or gives institutional weakness a technologically impressive appearance.
Pakistan’s educational crisis is not primarily a shortage of content. Textbooks, videos and online explanations already abound. The deeper shortages are capable teachers, functioning classrooms, reliable assessment and accountability. UNICEF estimates that 4.5 million children aged five to sixteen, one in three, are out of school in Khyber Pakhtunkhwa, with girls affected. No platform can teach a child who is not enrolled, provide safe transport to a girl whose family fears school, or retain a pupil whom poverty compels to work. Placing a digital teacher before children does not guarantee learning. Every technological promise rests on a chain: electricity, connectivity, devices, maintenance, technical support and an adult able to manage the classroom. If one link fails, the platform becomes another dark screen. Schools most in need of remote expertise may be least able to sustain that chain.
There is a fundamental difference between answering a question and understanding a learner. A human teacher notices hesitation, embarrassment, distraction and fear. She may discover that a child has memorised a definition without comprehending it, or that silence reflects language difficulty rather than ignorance. Teaching is not merely transmitting correct information; it is interpreting how children respond. Artificial intelligence may simulate conversation, but it cannot assume pastoral responsibility for a classroom. Its answers require supervision. Generative systems can produce confident inaccuracies, flatten contested questions and reproduce biases. The risks become especially sensitive in history, religion, citizenship and regional identity. Who approves the material? Which curriculum governs the answers? Are children’s interactions recorded, and who may access them? A textbook has authors, reviewers and an inspectable edition. An adaptive system may generate a different response for every pupil. Accountability must precede deployment, not follow controversy.
The government’s financial framing creates concern. Officials have projected that Parho Khyber Pakhtunkhwa would require about five billion rupees while generating savings exceeding one hundred billion. Such figures need a published calculation. If savings depend on leaving teaching posts vacant or replacing teachers with technical facilitators, technology will cease to supplement the system and begin shrinking it. To its credit, the Elementary and Secondary Education Foundation has recruited subject specialists and teaching staff for the Global Virtual School. That is the correct instinct. Yet recruitment for a pilot is not a permanent safeguard. Fiscal pressure can gradually recast vacant posts as opportunities for innovation and trained teachers as attendants responsible for operating screens and maintaining order. This reverses the proper relationship. Technology should extend a teacher’s reach, not furnish the state with a reason to withdraw the teacher.
The sounder model is teacher plus machine. An experienced educator can select an AI explanation, test whether pupils understood it, correct errors and connect an abstract lesson to local experience. The platform can supply practice questions, translation and simulation; the teacher supplies judgment, motivation and care. Technology then performs the tasks it does well without being assigned the human work it cannot carry. Before expanding the initiative, Khyber Pakhtunkhwa should conduct a carefully designed pilot across different realities: urban and rural schools, boys’ and girls’ schools, connected and poorly connected communities, and adequately staffed and understaffed classrooms. Baseline data should record attendance, teacher availability, electricity, connectivity and learning levels. Independent evaluators should then measure not how many schools received the platform or how many lessons were played, but whether participating pupils learned more than pupils taught without it.
The results should be public, together with the cost per student, procurement arrangements, data safeguards, procedures for correcting inaccurate content and training provided to classroom teachers. Pakistan’s education sector has accumulated too many projects measured by distribution rather than outcomes: devices supplied, laboratories announced, personnel trained and portals launched. The missing question is always the same: what changed for the child? Artificial intelligence will enter classrooms, and resisting it altogether would be futile and unwise. But urgency is not permission to avoid institutional design. Khyber Pakhtunkhwa can demonstrate not merely that an AI teacher speaks three languages, but that a public system can innovate without surrendering accountability. A machine may explain the lesson. It cannot carry the obligation Article 25-A places upon the state to provide free and compulsory education to every child between five and sixteen. That duty still belongs to governments, schools and human beings, and no software should be allowed to conceal their absence.
—The writer is Commoner from 44th Common Educationist — Founder of WHI Institute.based in Sargodha.
