The greatest risk of artificial intelligence is not that machines will outthink us. It is that they will inherit our oldest inequalities and write them into the infrastructure of the future, at a speed and scale no technology has ever matched.
Across the world, AI is already deciding who is shortlisted for a job, approved for a loan, prioritised for medical care or recommended for a scholarship. These decisions appear objective because a machine, not a person, makes them. But algorithms do not exist in isolation. They learn from data, and data is a mirror of the society that produced it. Where inequality exists in the real world, it is quietly written into the digital one.
For Pakistan, this is not a technology question. It is a development question, and how we answer it will shape the next decade of growth.
Pakistan is not watching this transformation from the sidelines. The country has surpassed 200 million telecom subscribers and 150 million broadband connections, while expanding digital public services, innovation centres, and technology exports. The Digital Pakistan vision and the National Artificial Intelligence Policy are positioning the country to compete as a serious digital economy.
The momentum is real. Pakistan’s IT and IT-enabled services exports reached a record $4.6 billion last fiscal year, marking a 21% increase over the previous year, with the sector expected to sustain this strong growth trajectory. AI can accelerate that trajectory: transforming agriculture, improving healthcare, modernising education and creating entirely new categories of economic opportunities. But momentum is not the same as inclusion, and growth is not the same as fairness.
Consider a young woman entrepreneur in Bahawalpur building an AI-enabled tool for precision farming, or a software graduate from Turbat applying for her first technology job. Now imagine how these challenges multiply for a young woman with a disability, a woman from a remote rural community, a member of a religious minority, or someone from a low-income household with limited digital access. Increasingly, their prospects will be shaped not only by people but also by algorithms that screen applications, assess creditworthiness, or recommend investments. When these systems are trained on historical patterns in which women and other marginalised groups had fewer opportunities to own businesses, build credit histories, access quality education or reach leadership, they risk reproducing and hard-coding those same inequalities.
AI is not designed to discriminate. It simply learns from a world that has not yet achieved equality. The same risk runs through generative AI. Ask a system to illustrate a chief executive, an engineer, or a scientist, and it is still more likely to generate a man. Ask it to portray a caregiver or a homemaker, and a woman often appears. Multiply that pattern by millions of daily interactions, and it stops being trivial; it becomes a quiet, constant lesson about who leads, who innovates, and who belongs in the digital economy.
Pakistan cannot afford to build a digital economy on analogue inequalities. Women are nearly half of Pakistan’s population, yet their labour force participation stands at around 24%, among the lowest in the region. The gap extends into the digital sphere, but here the story is changing. GSMA’s latest Mobile Gender Gap data shows Pakistan narrowed its mobile ownership gender gap from 37% to 27%, and its smartphone gap from 48% to 30%, in a single year: the sharpest improvement of any country surveyed. That is proof that targeted policy and industry action move the needle. It is also a measure of the distance still to travel.
Globally, women hold barely 30% of AI jobs; in Pakistan, far fewer women study and work in science, technology, engineering and mathematics, which means fewer women are designing the technologies that will shape all of our lives.
These are not merely social challenges; they are economic constraints. The IMF estimates that closing the gender gap could lift Pakistan’s GDP by as much as 30%. Every talented woman excluded from the digital economy represents innovation lost, productivity unrealised and growth deferred. Pakistan cannot aspire to regional digital leadership while leaving half its talent on the margins of the AI revolution.
Gender-responsive AI is not about building technology for women alone. It is about building technology that works fairly for everyone by recognising that people experience digital exclusion differently. Women and girls are not a homogeneous group. Rural women, women with disabilities, older women, young women entering the workforce, women from religious and ethnic minorities and women living in poverty often face multiple and intersecting barriers to accessing and benefiting from digital technologies. AI systems must therefore be designed with an intersectional lens that identifies and addresses these overlapping forms of discrimination rather than reinforcing them.
In practice, that means four things: diverse datasets, systematic bias testing, transparent governance, and design teams that reflect the societies AI is meant to serve. Get these right, and inclusive AI is not only more ethical but also more accurate, more trusted and ultimately more effective. This is the framework Pakistan should adopt, and the one we are now building into our own institutions.
That is why the Ministry of Information Technology and Telecommunication and UN Women, with the support of the Korea International Cooperation Agency (KOICA), are launching Pakistan’s first Gender-Responsive AI School. It will bring together policymakers, public institutions, academia, entrepreneurs, civil society and technology leaders to build practical AI skills grounded in fairness, accountability, and inclusion. Its purpose goes beyond training users: it is designed to ensure that women are among the innovators, researchers, founders and decision-makers shaping AI’s future, rather than simply consuming what others have built.
This matters because the world is still getting it wrong. The Global Digital Compact, adopted by world leaders in 2024, the agreed conclusions of the 67th Session of the Commission on the Status of Women, and the Beijing+30 agenda all recognise that digital transformation cannot succeed if it leaves half the population behind. Yet practice has not caught up with commitment: a UN Women review of 138 national AI strategies found that only 24 mention gender at all, and just 18 include concrete measures to address it. Pakistan now has the opportunity, and the institutions, to be counted among the minority that chooses a different path: one where innovation and inclusion advance together rather than in sequence.
Those choices must ensure that the benefits of AI extend to every woman and girl, including those who have historically been excluded because of disability, geography, poverty, age, ethnicity or other intersecting forms of disadvantage. A truly inclusive digital future is one in which AI expands opportunity rather than deepening existing inequalities.
Artificial intelligence will become part of the infrastructure of modern society, much as electricity and the internet did before it. The question is no longer whether Pakistan will adopt AI. It already has. The real question is whether we will build AI that reflects the values our constitution promises to everyone: equality, dignity and justice, or let those values become an afterthought.
History will not remember this generation for how quickly it adopted AI. It will remember whether we had the foresight to ensure that intelligence served everyone equally. If Pakistan gets this right, our greatest innovation will not be smarter technology but a fairer future — one in which opportunity is determined by talent and ambition, not by gender.
Shaza Fatima Khawaja is the federal minister for information technology and telecommunication, Government of Pakistan.
Jamshed M Kazi is the country representative for UN Women Pakistan.
Disclaimer: The viewpoints expressed in this piece are the writer’s own and don’t necessarily reflect Geo.tv’s editorial policy.
Originally published in The News
2026-07-29 09:45:00
