AI: A friend or a foe?

The past year has been marked by significant strides in AI, igniting debates about its implications for humanity. Is AI the universal solution we've been waiting for? Only time will tell. However, we must not merely observe but critically engage with these developments to truly understand their potential and pitfalls.

AI isn't a singular force but a powerhouse of capabilities—machine learning, natural language processing, robotics, and more—reshaping everything from healthcare to how we connect across continents. I've always seen technology as the great equaliser, breaking down barriers and democratising access to critical services. Yet, as we stand at this technological crossroads, we must confront a pressing question: Will AI continue to be a force for equity, or will it deepen the already prevalent digital divide?

Recently, at the World Economic Forum in Davos and MWC Barcelona, AI dominated discussions, emphasising the urgent need for robust governance frameworks. Yet, amidst this rush, we must be wary of falling into the 'sorcerer's apprentice' scenario—where our creation grows beyond our control. The narrative, however, is incomplete without addressing the potential for AI to be akin to the sorcerer's apprentice—powerful yet untamed. For developing nations like Pakistan, this power, if harnessed responsibly, could be a game-changer, but if mishandled, it could be a missed opportunity.

We are already seeing a range of AI-specific use cases being adopted across industries. From optimising network performance in the telecommunications sector to revolutionising fraud detection in fintech, AI's applications are proving their mettle. Yet, the journey from current use to widespread adoption is fraught with hurdles—infrastructure and skill development being paramount. Herein lies the potential for a robust AI strategy that champions inclusivity, leverages cost-effective innovation, and orchestrates skills tailored for our youth. Over the next few months, AI-led innovation in Agritech, Edtech, and Healthtech will be key in helping us address our critical challenges in these areas.

Yet, the fundamental challenge persists: How can Pakistan carve out its niche in the global AI landscape? It is critical that we undertake strategic and tactical initiatives without delay. Despite the excitement surrounding AI’s transformative capabilities, there is still a substantial gap in addressing the barriers to its widespread adoption in Pakistan—specifically, infrastructure and skills development.

Much has been discussed about AI's potential over the past year, yet there remains a lack of focused dialogue on the significant hurdles to its large-scale implementation in Pakistan. How can we unlock this enormous potential, and what critical steps are necessary today?

Foremost among these challenges is the inadequate infrastructure that hinders sustainable AI development. From limited broadband access to a shortage of high-compute data centres, both the industry and academia face considerable obstacles in developing cutting-edge solutions. The primary barrier to establishing the necessary infrastructure for AI is the substantial cost involved. AI research demands advanced computing clusters capable of producing solutions that meet global standards.

This leads to a massive skill development gap where our youth are dependent upon either low-end resources or compelled to explore opportunities outside the country to seek technological growth. The lack of a national vision to govern AI research results in a fragmented landscape that does not contribute to structured skill development. This not only affects the competitiveness of industries but also stunts the growth of a robust AI ecosystem within the country, exacerbating the brain drain phenomenon and further widening the skill gap. We also need to be mindful of our youth living in rural Pakistan. How can we build inclusive AI models that are all-encompassing?

We need to build a coalition around AI, which needs to draw from the Triple Helix Model of innovation, whereby key stakeholders including government, industry, and academia need to come together and drive AI forward. A key strategy would be developing an indigenous GenAI platform by building a local Large Language Model (LLM), which can be a critical step towards achieving a new level of national capability. Pakistan has around 77 established languages, and preserving this linguistic diversity should be a priority. By developing a local LLM, we may be able to improve accessibility and preserve local linguistic diversity along with developing our indigenous AI capability. And this is where AI could be a potential foe: while AI has huge benefits, one possible pitfall is the loss of languages, traditions, and even knowledge systems, which could be at risk of being overshadowed by dominant, prevalent global narratives and biases embedded in AI.

The shift from "Artificial" to "Augmented Intelligence" is essential. It's about augmenting, not replacing, human capacity with AI. This approach is vital for Pakistan's diverse society, ensuring that technology elevates everyone, and no one is left behind.