It is no longer a matter of choice: artificial intelligence (AI) has arrived. While the world acknowledges its transformative potential, even experts and business leaders in advanced economies are still grappling with its full implications and remain far from confident of managing them.
Pakistan, still at an early stage of digitisation with weak, fragmented data systems, appears largely unprepared. But a lack of readiness will neither shield the country from AI’s risks to its economy and labour market nor enable it to realise the dividends of technological advances in an inclusive manner.
A strange mix of scepticism and euphoria surrounds AI in Pakistan. Educated youth and a large percentage of service sector employees are familiar with and increasingly use ChatGPT and other AI tools, aided by widespread smartphone and internet access. Yet, with limited awareness and understanding of AI, many dread it will further erode already scarce job opportunities. At the same time, businesses are eagerly exploring cost-effective AI applications to reduce their wage bills while improving efficiency and productivity.
Dr Dur e Nayab, former director of research at the Pakistan Institute of Development Economics (PIDE), was candid about AI’s implications. “That’s a difficult question,” she said. “I don’t know anyone who truly understands AI’s full consequences, though that may also reflect my own limited knowledge of the subject”.
Part of the problem is that Pakistan’s statistical systems were designed for a pre-digital economy, long before personal computers, the internet and AI
Analysts familiar with AI remain cautious about its large-scale adoption in Pakistan’s public and private sectors, citing low levels of digitisation, poor data quality and the time lag in its availability and limited transparency. “AI is only as reliable as the data it is fed. Unless the government improves national data systems and businesses become more transparent in their accounts and operations, AI will neither deliver its full potential nor produce reliable optimal outcomes. Faulty data can lead to flawed diagnosis, poor decisions and public disruption,” an expert said.
Last week, Finance Minister Muhammad Aurangzeb announced that the government had introduced an AI-based tax administration model to curb corruption and improve efficiency.
A private sector expert said he did not question the government’s intent but doubted its institutional readiness. “The real problem is the limited understanding of AI across bureaucracy and political leadership. No matter how advanced, no AI model can deliver in a vacuum. Test political leaders across the aisle and federal secretaries on AI, and I would be pleasantly surprised if even five per cent passed,” he said.
Dr Naeem Zafar, Chief Statistician, Pakistan Bureau of Statistics, confirmed that the agency does not collect data on AI adoption in either public or private sector. While Pakistan Social and Living Standard Measurement (PLSM) Survey tracks the use of information communication technology (ICT), it contains no question on AI.
As a result, Pakistan lacks systematically compiled data on AI adoption across the economy. The provincial statistical bureaus remain practically dysfunctional for lack of focus and funds.
Part of the problem is that Pakistan’s statistical systems were designed for a pre-digital economy, long before personal computers, the internet and AI. For example, the Pakistan Bureau of Statistics’ labour survey tracks employment in broad sectors such as manufacturing, retail and construction, services, but not AI adoption across industries and businesses.
Nor do they identify occupations most vulnerable to automation, including software developers, editing staff, accountants, data crunchers, customer service agents, graphic designers, despite AI’s growing usage in banking, finance, telecommunication, advertising, media and other service industries.
According to available information, the Ministry of Information Technology has circulated a draft data protection and governance framework for public and industry feedback. It has also established the Pakistan Digital Authority and says it is developing technology and AI innovation hubs nationwide. Federal Minister for Information Technology and Telecommunication Shaza Fatima Khawaja was approached for an update and her views, but no response had been received by the time this report was filed.
Dr Sohail Munir, founding chairperson of the Pakistan Digital Authority, and head of the country’s digital transformation agenda, promised to share his views on the subject but later apologised, saying he was unable to respond before the deadline due to prior commitments.
Akif Saeed, former chairman of the Securities and Exchange Commission of Pakistan (SECP), said: “We still rely on photocopies and attested documents, so true digitisation remains incomplete. AI is being used only in limited pockets. At the company level, adoption varies widely, and the SECP does not maintain comprehensive data on AI usage,” he said.
Some experts argue Pakistan must rapidly adopt AI to boost productivity, lower costs and remain globally competitive, warning that delay will only widen the development gap. They caution, however, against blindly importing foreign models and stress the need for locally developed solutions tailored to Pakistan’s economic and social realities. Others believe the priority should be educating policymakers in the government and corporate boardrooms before AI is deployed at scale.
Muhammad Shahzar Illahi, Co-founder and CEO of EnablifyAI, said AI adoption in Pakistan remains limited despite widespread interest. Drawing on his firm’s work with UBL, insurers, Gallup, SECP, Telenor and Jazz, he said banks are leading adoption by developing in-house AI applications and hiring AI engineers, while most businesses lack a formal AI strategy. Although 80 to 85pc of executives and professionals they train use ChatGPT, many remain unaware of basic data security risks associated with public AI platforms.
He argued that business and government leaders need AI literacy before wider deployment. “AI is far bigger than ChatGPT,” he said, adding that poor data quality remains a major constraint. Mr Illahi warned that routine white-collar jobs, including data entry, financial analysis, compliance, human resources and other support functions, face the greatest disruption, while manual manufacturing jobs are harder to automate. He noted that the federal government has introduced AI and data governance policies, with Punjab leading provincial AI initiatives, but implementation remains at an early stage.
Asad Ali Shah was sceptical. “The government is busy copying policies, while the private sector shows little appetite for investing in new technologies. Innovation depends on young people, but we have failed to nurture the risk-taking mindset needed to achieve ambitious goals,” he said.
Published in Dawn, The Business and Finance Weekly, July 13th, 2026