he last two weeks saw two major AI summits in the region. The AI Impact Summit in India drew most of the attention as it was attended by world and industry leaders who flocked to New Delhi to be part of the new frontier of AI development. The Indus AI Week during which the government of Pakistan promised to invest $1 billion in AI by 2030 was less noticed. No details were provided about the form this investment will take, but these developments point towards how governments in the region view growth and the future.

Pakistan, too, seems to be hedging its bets on artificial intelligence to solve its complex economic problems. The optimism, however, falls flat given that many of these conversations skirt around serious governance, human rights and climate-related issues associated with AI. Further, the economic and growth-centric narrative around AI casts it as the magic wand that will gloss over the real structural problems that exist in our societies such as inequality, human rights backsliding and environmental degradation.

During the AI Week, the Ministry of Information Technology and Telecommunication announced the Islamabad AI Declaration, which it described as “a principled framework focused on public value, trusted data stewardship, accountable systems and global leadership in responsible AI.” The document outlines nine principles: sovereignty, human accountability, use-case-first, explainable systems, whole government governance, inclusive and responsible innovation, private-sector-led compute; and global leadership and AI diplomacy. Key amongst these principles is the issue of sovereignty, one that Pakistan has been attempting to legislate on for a while, particularly with reference to data sovereignty.

Several iterations of the personal data protection law have contained data localisation provisions, which capture the idea that local data, i.e. data of citizens and local businesses, should remain in Pakistan. While digital sovereignty sounds like a viable approach to AI, the idea of state consolidating personal data within its larger surveillance network is poised to weaken the sovereignty of individuals over data.

It is encouraging that the declaration also refers to accountability mechanisms and auditable systems within the framework of constitutional authority. Grounding AI in principles of transparent systems and the constitutional framework is important, particularly the human rights chapter in Pakistan’s constitution. Any move towards an AI future that ignores these principles must be deemed anti-people. While this language in the declaration and references to AI ethics in the 2025 National AI Policy are welcome, there exists no robust framework to ensure actual accountability. Viewing AI as mere economic good and lever for growth discourages accountability and makes the rollout of AI beholden to corporate interests.

The Pakistan Digital Authority has been tasked with operationalising some of these principles, but there is little clarity on how it will function as an independent watchdog rather than a bureaucratic extension of the state.

It is also worth noting that Pakistan’s digital surveillance and datafication push sits uncomfortably with its AI ambitions. The lack of privacy safeguards and an unchecked consolidation and centralisation of digital data in the name of efficiency and better service delivery are a cause for concern. Introducing AI into the mix is the next logical step, but no thought is given to the implications of deploying AI in a fundamentally unequal society. How will AI-driven public infrastructure impact marginalised communities who already sit outside the digital world, either falling through the cracks of Pakistan’s yawning digital divides or being pushed to increasingly give up data to AI-addled, data-hungry systems?

There are obvious tensions between the need for inclusion into these systems and the broader implications of these systems for the world at large. These tensions remain unresolved in Pakistan’s vision for an AI-enabled future.

One of the most glaring omissions in Pakistan’s AI discourse is its environmental cost. The AI industry has been marked by climate change experts as dangerous for the environment, particularly its massive strain on water resources used for cooling. As an already water-scarce country experiencing unprecedented glacial melt, any discussion on AI in Pakistan must answer the question: are the gains made by AI offset by its outsized environmental costs? If our vision for AI does not take into account the future of our planet, then we are hurtling towards an unsustainable future. In our race towards an AI economy, climate change should not be treated as a party spoiler. Any discussion, declaration or governance framework about AI is incomplete without a climate change lens.

Pakistan is at a crossroads; it can choose to blindly pursue AI as a driver for economic growth or build a future that places the needs of its people and society at the centre of any technological growth. AI can be a facilitator and tool for making our lives better in the future, but we cannot let it determine the values, rights and priorities that shape our collective future.

The writer is a researcher and campaigner on human and digital rights issues.