Doctrine, Policy, and Strategy

NAYA PAKISTAN, PURANA SOFTWARE: AI MEETS PAKISTANI REALITY

If the word on the street is to be believed, Artificial Intelligence (AI) has arrived to rescue Pakistan from problems that generations of human intelligence have failed to solve. It can prepare economic plans, improve tax collection, detect corruption, and explain why a country blessed with extraordinary talent and entrepreneurial energy continues to negotiate with the International Monetary Fund! Your guess is as good as mine! AI can produce a national development strategy in seconds. Pakistan, meanwhile, has spent decades producing strategies that demonstrate a remarkable ability to survive without being implemented. No small feat this!

Imagine if you will the launch of a National Artificial Intelligence Transformation Initiative. The ceremony would take place in a five-star Islamabad hotel, complete with an impressive LED screen, a lavish buffet, and enough official photographs to fill an annual report. A federal minister would announce that Pakistan was on the verge of becoming a global AI powerhouse. Consultants would present a 200-page strategy, and a committee would be formed to oversee its implementation. Six months later, the committee would recommend another committee to assess progress. AI would have completed the entire exercise in seconds. Unfortunately, the ministry’s website would be down, procurement would take eleven months, and the project would be postponed until the next financial year. Pakistan would have successfully launched artificial intelligence without actually requiring any intelligence to be deployed.

“The machine” might also be confused by the country’s economic policies. It would analyse government spending, tax revenues, public debt, and repeated promises of economic recovery. It might conclude that Pakistan’s most successful industry was the production of economic recovery plans. Every few years, a new government announces that the country is standing at a historic turning point. The previous turning point, apparently, was facing the wrong direction!

AI would recommend broadening the tax base, improving public-sector efficiency, encouraging investment, and reducing waste. These suggestions would be so sensible that they might cause discomfort in certain official circles. After all, why would anyone need artificial intelligence to explain that a country cannot indefinitely finance its ambitions through borrowing, taxing the already compliant, and hoping that the next IMF programme will somehow be the last?!

Corruption would offer another fascinating challenge. AI could recommend transparent procurement, digital records, automated audits, and public dashboards showing where taxpayers’ money actually goes. It might even flag suspicious discrepancies between a project’s approved budget and its final cost. The technology would be useful, but it would have one significant limitation: it could identify irregularities without necessarily ensuring that anyone investigated them. An algorithm can flag a questionable transaction. It cannot compel an institution to pursue the matter. It can recommend accountability, but it cannot make accountability convenient for those who prefer to avoid it. In Pakistan, the most revolutionary feature of an AI system might therefore be its memory! It could remind citizens that a project announced as nearly complete several years ago is still awaiting completion. Imagine a machine that remembers every government promise, compares it with actual performance, and sends polite reminders whenever deadlines are missed. If sufficiently advanced, it might even learn to copy the reminders to senior officials.

The private sector would embrace AI with equal enthusiasm. Corporate presentations across Karachi, Lahore, and Islamabad would suddenly feature phrases such as “AI-driven transformation”, “unlocking synergies”, and “exponential growth”. A company struggling to maintain accurate inventory records would announce that it was entering the fourth industrial revolution. Its chief executive would present a five-year forecast showing profits rising confidently towards the heavens. The finance director would quietly wonder how the company intended to finance its next shipment.

Consultants would flourish, selling AI-readiness assessments, digital transformation roadmaps, and strategic implementation frameworks. Businesses could spend millions learning how to adopt AI without deciding which problem they wanted it to solve. The final presentation would contain 64 slides, four strategic pillars, and several overlapping circles representing people, processes, and technology. Nobody would understand what the circles meant, but everyone would agree that they looked impressive.

Meanwhile, Pakistan’s small businesses would face more practical questions. A textile manufacturer in Faisalabad might use AI to reduce waste and improve production planning. A trader in Karachi could forecast demand and manage inventory. A farmer in rural Punjab or Sindh might benefit from better crop advice and weather information. These are genuine opportunities, but they require affordable technology, reliable electricity, relevant local data, and appropriate training.

Employment presents another challenge. Pakistan’s young population offers enormous potential, and AI could support education, entrepreneurship, and new industries. But automation may also disrupt routine jobs before alternative opportunities become available. A minister might announce that AI will transform Pakistan into a global technology powerhouse. A young graduate with a degree and a growing collection of unanswered job applications might reasonably ask how. The minister would recommend digital skills. AI would generate a comprehensive training curriculum. The graduate would complete the course and discover that every advertised entry-level position requires three years of experience.

Perhaps AI’s greatest weakness is its ability to reproduce human mistakes at extraordinary speed while wrapping them in impeccable grammar. It can generate a development policy without understanding local realities, a business plan without appreciating financial constraints, and a corruption-prevention framework without recognising the institutional incentives that allow corruption to persist. It can recommend meritocracy to a system accustomed to patronage and long-term planning to leaders whose attention is firmly fixed on the next election or quarterly result. Yet the machine is not entirely to blame. AI reflects the information, assumptions, and priorities supplied by the people who build and use it. If its recommendations are superficial, we should examine the questions. If governments use it to produce impressive announcements rather than measurable improvements, we should not be surprised when it becomes another expensive monument to unfulfilled ambition.