Sri Lanka’s AI adoption trails globe and region despite hub ambitions - Microsoft report

By Nishel Fernando

Sri Lanka’s adoption of artificial intelligence lags significantly on the global stage, placing it among the nations with the lowest AI usage rates, a major news report has revealed. 

The country’s AI diffusion rate stands at a mere 6.2 percent, a figure that not only trails far behind global leaders but also falls below the average for the Global South and many of its South Asian neighbours.

This stark reality, detailed in the November 2025 ‘AI Diffusion Report’ from the Microsoft AI Economy Institute, presents a significant challenge to the nation’s declared ambition of becoming a regional AI hub. The report highlights a critical gap between the current state of AI integration and the strategic goals necessary to harness the technology’s economic potential.

The Microsoft report provides a comprehensive overview of global AI adoption, which it terms “diffusion”- the process by which technology becomes integrated into society. It notes that while AI is the fastest-spreading technology in history, with over 1.2 billion users in under three years, its benefits are not being shared equally.

A key finding is the emergence of a stark “North-South Divide.” AI adoption in the Global North is approximately 23 percent, nearly double the 13 percent average in the Global South. The report identifies several foundational “building blocks” that are critical for AI adoption, including reliable electricity, internet access, data centre infrastructure, and digital skills. Nearly four billion people, almost half the world’s population, still lack these basic prerequisites to use AI.

Language has also emerged as a powerful new barrier, with countries dominated by “low-resource languages”—those with less data available to train AI models—showing significantly lower adoption rates.

Sri Lanka’s 6.2 percent AI diffusion rate places it in the bottom tier of the 100 plus economies surveyed. This performance is a fraction of that seen in leading nations like the United Arab Emirates (59.4 percent) and Singapore (58.6 percent), which have demonstrated that strong digital infrastructure and policy coordination can drive rapid adoption.

The comparison with regional peers is particularly telling. Sri Lanka lags considerably behind India (14.2 percent), Nepal (12.3 percent), and Pakistan (9.7 percent), and even trails Bangladesh (6.5 percent). Its adoption rate is also dwarfed by Southeast Asian competitors such as Vietnam (21.2) and the Philippines (17.1 percent), nations that often compete for similar technology and outsourcing investments. This places Sri Lanka at a competitive disadvantage in a region that is rapidly embracing digital transformation.

The sluggish adoption rate contrasts sharply with Sri Lanka’s strategic goals. The government has been formulating a national strategy for Artificial Intelligence, aiming to position the country as a hub for AI innovation and high-tech talent. This strategy focuses on developing human capital, attracting investment, and promoting the use of AI in key sectors of the economy.

However, the Microsoft report suggests that fundamental barriers may be hindering this vision. For a country to successfully build an AI ecosystem, it must first ensure the foundational layers are in place. The report’s findings imply that Sri Lanka must urgently address these core challenges if its AI hub strategy is to become a reality. Potential barriers include gaps in digital literacy, inconsistent access to high-speed internet, and a need for greater investment in the underlying digital infrastructure required to power AI applications at scale.

A significant hurdle, as highlighted by the report’s global analysis, is language. While the report does not mention Sinhala specifically, the language has historically been considered low-resource in AI development. Encouragingly, this is beginning to change. Researchers at the University of Moratuwa recently developed ‘SinLlama,’ the first open-source LLM with explicit support for Sinhala, by adapting Meta’s Llama-3 model. 

This is along with commercial efforts by firms such as Sigiri. AI, represents a crucial step in making AI more accessible and relevant to the local population, directly tackling the language barrier identified in the report.