Pakistan’s AI reckoning
0
PAKISTAN is beginning to confront a question it can no longer defer: how to AI-proof its future. Artificial intelligence is no longer a distant ambition. It is an immediate necessity. In recent months, Pakistan has announced a nationwide AI training programme, committed $1 billion in investment by 2030, and outlined plans to train a large segment of its workforce in AI-related skills.
These commitments, reflected in the Islamabad AI Declaration unveiled during Indus AI Week 2026, signal that Pakistan is no longer merely observing the AI transition from the sidelines. Delay now carries consequences the country can no longer afford.
That recognition, however, is only the starting point. Pakistan’s response remains primarily at the level of intent. AI, on the other hand, rewards preparation, speed and continuity. Intent has rarely been Pakistan’s problem. It has been execution. Announcements create momentum, but only sustained implementation creates outcomes.
Pakistan does not suffer from a shortage of ambition. It suffers from a shortage of follow-through. And it is precisely at that point, between announcement and execution, that momentum in Pakistan has historically thinned out.
Pakistan has approximately 60 per cent of its population under 30, but this is not an advantage unless its youth are equipped and trained. Millions remain outside the education system, and many within it lack usable skills while at the same time, entry-level jobs are being automated or redefined.
The global conversation surrounding AI has already shifted. Unesco insists on embedding AI literacy across education systems. The ILO warns about large-scale reskilling. UNDP emphasises a widening capability gap. Taken together, these are not isolated concerns but signals of a shared direction. Without sustained intervention, inequality is likely to deepen. This is not just a technological transition. It is a reordering of skills and opportunity, and its effects will not be evenly distributed.
Pakistan needs to ensure that AI is integrated into the national ecosystem.
To its credit, Pakistan is no longer absent from this conversation. Its National Artificial Intelligence Policy Framework approved by the federal cabinet in July 2025 signals initiative. However, a closer look reveals structural weaknesses. One assessment described the policy as ambitious but lacking legal teeth and overly reliant on optimism. In effect, Pakistan’s framework gestures in the right direction but stops short of creating binding obligations.
Another gap emerges in Pakistan’s legal and regulatory landscape. The Prevention of Electronic Crimes Act provides a baseline for cyber offences, while the proposed Personal Data Protection Bill remains pending. The State Bank of Pakistan’s digital banking regulations cover some automated processes such as fraud detection and onboarding, but the underlying legal foundation dates back to 2007, well before generative AI. This leaves unanswered questions around liability and accountability. These are not peripheral. They sit at the core of how AI systems will operate.
Then there is the question of what to do with the bodies already created to manage AI such as the National Centre of Artificial Intelligence, set up in 2018, which functions as a hub for research, training, and knowledge transfer.
However, as a non-regulatory institutional initiative, it does not possess enforcement powers. More recently, institutional mechanisms introduced under the National Artificial Intelligence Policy 2025, including proposed coordination and oversight structures, have begun to emerge, but their regulatory authority and practical impact remain limited and are still evolving.
Questions of liability further complicate matters. Pakistan’s legal system rests on actus reus (guilty act) and mens rea (guilty mind). Autonomous systems disrupt both. When an AI system produces harm, the act is identifiable, but the mind behind it is not. The question then becomes who bears liability. The developer, the deployer, or the institution relying on the system? Pakistan’s laws offer no coherent answer. Without clear standards, business innovation slows and responsibility diffuses.
Education — and by extension AI education — is now a provincial subject after the 18th Amendment to the Constitution. Provincial governments control curricula and training, while the federal role is largely coordination. This makes alignment essential, yet current efforts remain fragmented. What is missing is a mechanism that requires coordination rather than merely encouraging it. Without it, even well-designed initiatives risk working at cross purposes.
In addition to resolving legal and regulatory gaps, Pakistan needs to ensure that AI is integrated into the national ecosystem.
Skills training must lead to demonstrable ability, not just certification. Final-year students should complete industry projects, and a portion of course credits should come from laboratory work. The government should also publish an annual skills gap analysis mapping curricula to employer needs.
The private sector must also be integrated within the state’s efforts. Companies cannot remain passive consumers of trained talent. The state should offer tax credits for certified apprenticeships, create a national internship platform, and require firms bidding for public contracts to demonstrate AI training pipelines. Without this linkage, training and employment will remain disconnected.
Access must also be addressed. Without deliberate effort, AI’s benefits will remain concentrated in major cities. Equipment grants should prioritise rural institutions, and training programmes should reserve seats for underserved districts. If access follows existing infrastructure, the divide will only widen.
Execution is not about a single reform and ultimately depends on alignment across education, industry, infrastructure, and law. If one lags, the rest will not hold. The challenge is systemic. It demands coordination, discipline, and sustained political will, not isolated bursts of effort or short-lived initiatives.
AI will not wait for Pakistan’s readiness, nor will it adjust to its pace. It will move forward regardless, redistributing advantage in real time. The question is whether Pakistan moves with it, deliberately and with conviction, or watches from the sidelines. That choice is no longer theoretical, and it is no longer comfortably distant. It is already unfolding. And the consequences of getting it wrong will not be gradual. They will be decisive, felt across generations, and exceedingly difficult to reverse.
The writer is an international law practitioner and a graduate of Harvard Law School.
Published in Dawn, May 4th, 2026
