Getting the data right

EDITORIAL: The growing use of artificial intelligence in economic planning, public administration, and service delivery has brought an old problem into sharper focus: technology can be only as effective as the data it relies on.

For Pakistan, where statistical information remains fragmented across government institutions and is not always collected according to consistent standards, the challenge is to ensure that the push towards digitalisation is matched by a commitment to data integrity.

Discussions at the recently held ‘AI Tech Expo’ in Islamabad underscored why this must become a national priority. A key concern raised at the event was the need to align Pakistan’s statistical systems with internationally recognised frameworks, including those used by the IMF and the World Bank.

Pakistan Bureau of Statistics Director Rizwan Bashir advocated adopting the Statistical Data and Metadata exchange, a standard that facilitates the consistent sharing and comparison of statistical information. Such alignment would make national datasets more accessible and useful, both for Pakistani policymakers and for international economic analysis. More importantly, it would help address the incompatibility between datasets produced by different public institutions.

This is a fundamental weakness in Pakistan’s statistical architecture. The PBS, provincial statistical bodies, ministries, and other public agencies collect information for different purposes, often using varying methodologies, definitions and reporting practices. Inadequate resources, limited technical expertise and weak coordination further compromise consistency.

The resulting discrepancies can distort key economic indicators, from per capita income to sectoral output, undermining economic assessments, resource allocation and the evaluation of public spending. Without comparable and verifiable figures, policymakers risk basing decisions on an inaccurate understanding of the country’s economic realities.

The consequences, however, extend beyond economic measurement to the everyday distribution of public services and resources. This was illustrated by the country’s former chief statistician at the event, who highlighted the potential of Pakistan’s digital population census, noting that it had resulted in the geocoding of over 40 million structures.

By integrating this information with administrative records, economic surveys and geographical datasets, policymakers could develop a more precise understanding of where public services are needed most. Mapping school-age populations against the location of educational facilities, for instance, could help identify where new schools are required.

Similar applications could inform healthcare provision, infrastructure development and investment planning. Realising this potential, however, depends on the accuracy of the underlying data and effective coordination among institutions to make their records accessible and compatible.

The implications are particularly significant as AI becomes more deeply embedded in decision-making, making it essential that more attention is directed towards data quality and reliability. Algorithms trained on incomplete, outdated, or inconsistent information can reproduce errors and generate misleading conclusions, giving flawed analysis the appearance of scientific precision.

AI cannot compensate for defective statistical foundations. If anything, it makes rigorous verification and standardisation more urgent, as errors can be replicated across systems and influence decisions at scale. Poor data integrity also limits the practical value of Pakistan’s technological advances, forcing businesses to rely on foreign datasets that may not reflect local realities. Making reliable domestic data accessible is therefore essential.

What we need to address these shortcomings is firstly the harmonisation of statistical standards across federal and provincial institutions, supported by transparent quality control procedures, regular audits and safeguards against selective reporting.

Equally necessary is a larger cadre of trained statisticians, data scientists and analysts within government departments, capable of both protecting the integrity of information and translating it into practical policy insights. Universities, public agencies and industry must collaborate to build these skills.

Pakistan’s technological ambitions must rest on a dependable statistical foundation. Standardised, credible and accessible data would strengthen economic planning, improve public services and make emerging technologies more useful.

Copyright Business Recorder, 2026