ISLAMABAD: Vague commitments and structural blindspots in the policy risk undermine both public trust and equitable access to emerging technologies.

This was said by participants of a multi-stakeholder dialogue on Pakistan’s recently approved National Artificial Intelligence (AI) Policy.

The dialogue was organised by the Institute for Research, Advocacy and Development (IRADA) in Islamabad, moderated by IRADA Advocacy Lead Moaz Bhangu.

The participants said while the policy was structured around six pillars and outlined ambitious goals under its “Awareness and readiness”, its pledges on privacy, security, education, and inclusion remain vague, with no clarity on how a safe and fair data governance framework will be ensured. Without strong provisions, they argued, women, marginalised groups and smaller institutions could face further exclusion from the benefits of AI.

They said the absence of comprehensive data governance will leave a dangerous gap at a time when data protection, privacy, and accountability were critical for building public trust in AI.

Humaira Mufti from the National Commission on the Status of Women (NCSW), while discussing policy’s gender dimensions, warned that despite rhetoric about inclusivity, the policy does not adequately address capacity building for women or protection against gendered disinformation.

Irfana Yasser from BBC argued that women risk being further marginalised if targeted training and safeguards were not concretely built into implementation plans.

Hamza Khan of Unesco added that Pakistan’s Right to Information (RTI) laws also needed rethinking in the light of AI. He stressed that without stronger transparency mechanisms, citizens will have little recourse to challenge misuse of their data or AI-driven decisions.

Similarly, think-tank Tabadlab cautioned that Pakistan’s challenge was not technical capacity alone but behavioural and institutional barriers. “The problem is not tech. The problem is integrating and using tech,” one representative said, pointing out that the ecosystem must be designed to encourage ethical use and to build confidence among citizens rather than fear.

Haroon Rashid of Independent Urdu noted that the country’s AI sector was still nascent, while the policy appeared overly focused on the public sector.

Dr Saadia Ishtiaq Nauman, Assistant Professor at Fatima Women University (FJWU), said Pakistan’s universities were already well connected to the diaspora, where Pakistan-origin technologists were engaged in groundbreaking work.

Ayesha Arif, an AI engineer, agreed that while tools can be built, the success of AI in Pakistan will depend on whether solutions were context-sensitive and rooted in audience needs.

“We have to start somewhere,” she said, “but we must start with the right questions.”

Moaz Bhangu of IRADA cautioned that while the policy rightly aspired to build indigenous AI solutions, its gaps on localisation and language barriers risk turning Pakistan into a training ground for exportable foreign skills.

Published in Dawn, August 24th, 2025