Sri Lanka has the luxury of not being spoiled for choice when it comes to participating in the global artificial intelligence (AI) boom. While it may not have, for example, India’s technological infrastructure, it is also not burdened by the competing interests of larger populations. As such, a strategic and focused approach could benefit Sri Lanka more than an attempt to follow in the shadow of other nations.
Dr Moinul Zaber, a data and computational social scientist who focuses on Artificial Intelligence and Telecommunication Technology for public policy, believes that low-income countries like Sri Lanka can get started immediately with relatively little effort. “For instance,” he points out, “encouraging the use of open-source AI tools and frameworks allows countries to adopt cost-effective, transparent, and customizable solutions while reducing reliance on proprietary technologies. Collaborating on open data initiatives further diminishes dependency on external providers, enabling local innovation and the creation of systems that address societal challenges.”
He is optimistic about AI’s future in the policy framework of developing countries. “What excites me about this field,” he says, “is its potential to reduce uncertainty in complex decision-making and provide actionable insights for societal challenges. From urban infrastructure planning to discovering and addressing disparities in social security systems, AI offers tools to empower people and promote equitable development. My work aims to make this promise a reality while ensuring ethical and responsible AI usage.”
British-based Tortoise Media has run a Global Artificial Intelligence Index (GAII) since 2019, measuring ranking countries based on their capacity for artificial intelligence by measuring levels of investment, innovation, and implementation. In 2024, the fifth edition of the GAII, Sri Lanka was ranked at 82/83, beating only Ethiopia in terms of AI-ready talent, infrastructure, and even government strategy, among other metrics. India comes in 10th, though it notably ranks 2nd in talent, conceding the top spot only to the US (which dominates the table).
Similarly, the Oxford Insights Government AI Readiness Index 2024, which specifically considers how ready governments are to implement AI in the delivery of public services, ranks Sri Lanka 85th out of 188 countries. It is clear that Sri Lanka cannot pursue every AI avenue possible with its limited resources; therefore, it should focus on implementations that are most likely to work as expected. The cultivation of homegrown AI talent to tackle larger challenges rests somewhere beyond the horizon, but the results of everyday AI solutions can start flowing in today.
Where Can AI Make an Impact in Public Policy?
According to Dr Zaber, AI stands to offer the most benefit in decision support systems, disparity identification, predictive analytics, and social security systems. In the first, machine learning can be used to analyse large datasets, providing actionable insights for urban planning and resource allocation. AI can also identify disparities by integrating satellite, sensor, and socioeconomic data to reveal hidden patterns of inequality, helping governments target resources effectively. On the other hand, predictive analytics tools enable proactive action by forecasting trends like economic changes or climate-related risks. Finally, AI can enhance existing social security systems by identifying coverage gaps and vulnerable populations, personalizing services to ensure comprehensive support for all.
This notably has relatively little to do with generative AI, which is as controversial in some spheres as it is a darling in others. “The practical difference between generative AI and machine learning lies in their focus,” he explained. “Machine learning is a broader domain where systems learn patterns from data to make predictions or decisions. Generative AI, a subset of machine learning, creates new content—be it text, images, or designs—based on learned patterns. While generative AI excels in creative and exploratory tasks, machine learning has a wider scope in prediction, classification, and optimization problems.”
Examples of this are already making waves around the world. Uganda, known for being one of the world’s poorest and least-developed countries, nonetheless boasts excellent mobile digital penetration. The nonprofit BarefootLaw capitalized on this by launching Winnie, an AI lawyer that provides pro-bono legal services to Ugandans, covering nearly a million users and resolving around 20,000 cases to date. While there are concerns about such services being accurate enough for deployment in this manner, BarefootLaw’s example nonetheless represents an improvement to many who once had no legal support whatsoever.
In Togo, the pandemic made it unfeasible for the government to manually locate and register people in need of social assistance, leaving its most vulnerable populations without aid. In response, a fully digital system, NOVISSI, was implemented, utilizing government administrative data sources as well as satellite imagery, data science, and machine learning methods to prioritize those most affected by the crisis. Roughly a year after NOVISSI was launched in April 2020, it had successfully disbursed assistance to nearly 1 million people, a majority of whom were women who used it to support their households.
Like Uganda and Togo, Sri Lanka has government programmes that are intended to support and protect its most vulnerable populations (among others). Likewise, Sri Lanka has excellent mobile penetration. It similarly follows that Sri Lanka has opportunities to use AI to improve the effectiveness of public policy.
Laying the Foundation
“Governments should focus on capacity building as a cornerstone of their AI strategies,” says Dr Zaber. “Developing training programmes, establishing university-industry partnerships, and offering scholarships can create a sustainable pipeline of AI talent. Collaborations with international organizations and academia can further enhance access to cutting-edge knowledge and technologies. By investing in education and infrastructure, governments can empower their citizens to lead AI innovation and implementation.”
According to the Committee on Formulating a Strategy for Artificial Intelligence (CFSAI), three major initiatives have already been launched to this effect. Artificial Intelligence in Sri Lanka (CFSAI, 2024) describes the formation of a Presidential Secretariat-led multi-stakeholder committee to craft a National AI Strategy for the period 2024-2028, an investment of Rs1.5 billion from the 2024 budget to lay the groundwork for AI, and the establishment of the National AI Center to direct AI implementation across the country.
The White Paper also lists certain challenges, such as a migration-induced brain drain following the recent economic crisis; middling basic computer, digital, and data literacy; incomplete digitalization; and so on, many of which are already being addressed. For instance, a connected digital government is one of the goals in Sri Lanka’s Digital Strategy 2030, and this is crucial given that AI integration is considered most effective when digitization and digitalization of the wider economy are already present. Similarly, data and AI literacy, from primary to tertiary education, is listed among the long-term goals for Sri Lanka’s AI Vision.
Dr Zaber’s perspective aligns with this emphasis on education. He says, “A critical aspect of contextualizing AI is empowering universities across all disciplines, not just computer science departments. Universities are centers of knowledge creation and innovation, and involving humanities, social sciences, and engineering departments in AI research ensures a multidisciplinary approach to designing systems that address diverse societal needs. For instance, Finland has integrated AI education into non-technical fields by offering the online course ‘Elements of AI,’ which has been adopted by people from various disciplines, including arts and humanities. This initiative demonstrates how AI education can be democratized and contextualized for broader societal impact.”
Homegrown Solutions & Big Tech
Sri Lanka faces the unenviable challenge of deciding how to broadly implement artificial intelligence (AI) across its public services without running the risk of foreign dependence or widening the technological divide. Already notorious for playing catch-up in the digitalization of its public services, the country now has to reckon with a technology whose core demands include vast amounts of digitally available data, yet it can ill-afford to proceed too cautiously lest it falls behind for the next decade and beyond. On the other hand, it needs to remain wary of investing too recklessly in an industry that some, such as former Stability AI CEO Emod Mostaque, are warning bears similarities to the dot-com bubble of the late 90s.
Artificial Intelligence in Sri Lanka separately notes that few AI solutions were in development at the time of publication, either in the public or private sectors. Instead, local AI usage was, and perhaps still is, represented almost exclusively in ChatGPT, Microsoft Copilot, Google Gemini, and so on. However, this does not mean that the country is obliged to compete with these products. Instead, it is advised that they be utilized to enhance workflow and productivity—and local solutions can be developed that likewise address local needs.
“While over-reliance on Big Tech is concerning,” Dr Zaber comments, “initiatives such as Google’s AutoML, Facebook’s PyTorch, and Microsoft’s AI for Good and AI for Earth programmes offer valuable resources that can be strategically leveraged. AutoML, for example, simplifies machine learning model development, making AI more accessible to non-experts. Similarly, PyTorch and fairness toolkits enable local developers to create ethical and transparent AI solutions. Microsoft’s programmes support projects in critical areas such as sustainability, agriculture, and education, providing tools and funding that can accelerate localized innovation. These initiatives, combined with an open data culture, democratize access to AI technologies and reduce barriers for local stakeholders.”
With Big Tech and affluent countries vying for influence in the AI sphere at all levels, it is important for developing countries to insulate themselves from undue dependence, bias, or interference, intentional or otherwise. Dr Zaber argues that AI strategies should focus on inclusivity, open data, and sovereignty. Investing in infrastructure like national data centers, for instance, would safeguard data sovereignty and reduce dependency on external platforms. While independent regulatory bodies may provide oversight to ensure ethical use, he says governments must support local innovation with funding, tax incentives, and startup assistance, again limiting how often the country has to turn to solutions that may have been developed for different audiences and environments.
On the other hand, Sri Lanka need not tackle this task alone. By building AI infrastructure alongside private entities, local industries can be developed into viable alternatives to international providers. Going further, regional collaboration can reduce the cost of entry while keeping solutions based as close to home as possible. Dr Zaber suggests alliances with regional blocs, such as the Association of Southeast Asian Nations (ASEAN), to pool resources, share expertise, and more. Certain areas may also benefit specifically from this unified approach, such as agricultural optimization and disaster management, reducing the individual investment for each country while improving overall effectiveness.
The 2024 Draft Strategy for Public Consultation on Sri Lanka’s National AI Strategy notes the following: “In the wake of the economic crisis, harnessing the power of AI is not merely an option, but a necessity for Sri Lanka’s future. By integrating AI-driven solutions into public service delivery, optimizing resource allocation, reducing costs, and improving the overall effectiveness of government services, we can navigate the challenges posed by the economic crisis and lay the foundation for a more resilient, responsive, and citizen-centric government in the long run.”
Dr Zaber is optimistic that AI shows promise for developing countries like Sri Lanka but points out that the path forward must be navigated carefully. Despite the potential for AI’s integration into public services, the government needs to seriously consider feasibility, equity, and long-term sustainability at every step. Only a conscientious implementation of this technology will ensure that it genuinely serves the public good instead of becoming yet another footnote in the background of economic woes.