As the global artificial intelligence arena accelerates into a capital-intensive race defined by trillion-parameter models, billions of dollars in investments, and massive data facilities, Dr. Mothilal de Silva, Chairman of the SLT Group, has called on Sri Lanka to chart a distinctly different trajectory. Speaking in Colombo on the future of artificial intelligence in Sri Lanka, Dr. de Silva urged policymakers and industry leaders to stop chasing sheer scale and instead focus on high-impact, task-specific systems capable of solving practical domestic challenges. He emphasized that the next frontier of artificial intelligence will not necessarily belong to the largest models, but to smarter and more resource-efficient applications.
The push comes amid escalating discussions across Asia and internationally regarding artificial intelligence governance, where concerns surrounding algorithmic bias, safety, privacy, and accountability are transitioning from theoretical debates into regulatory imperatives. Dr. de Silva noted that artificial intelligence has shifted from being purely a technological breakthrough to an urgent governance challenge, highlighting critical questions about accountability when autonomous agents cause harm, methods to eliminate bias, and strategies to protect citizen data while fostering technological innovation.
Addressing the domestic landscape, Dr. de Silva identified computational infrastructure as the nation’s most critical gap, pointing out that Sri Lanka currently lacks the GPU capacity, hyperscale data facilities, and computational resources required to train massive language models on the scale observed in the United States, China, or Europe. However, tangible initiatives are progressing to bridge this divide. The Government has committed billions of rupees toward national AI and digital infrastructure, formal discussions are underway with global hyperscale tech firms, and SLT-Mobitel is actively preparing to upgrade its data centres to be AI-ready. Furthermore, Sri Lanka has already deployed more than 70,000 kilometres of fibre infrastructure, with high-speed connectivity projected to reach every secondary school nationwide by the end of the year.
Key Developments
- Adopting a ‘Small is Beautiful’ Strategy: Moving away from massive architectures like GPT-3’s 175 billion parameters and newer trillion-parameter models, Sri Lanka is encouraged to deploy smaller, specialized language models that offer superior task performance at a fraction of the computational and financial cost.
- Prioritizing Localized Solutions: Key focus areas include deploying AI assistants fluent in Sinhala and Tamil to streamline public administration, rolling out offline-capable smartphone apps for farmers to diagnose crop diseases, integrating diagnostic tools in rural healthcare clinics, and delivering tailored educational platforms to secondary students.
- Strengthening National Digital Infrastructure: Ongoing national projects feature billions of rupees in state funding, data centre modernization by SLT-Mobitel, global hyperscale partnerships, and the expansion of over 70,000 kilometres of fibre network across Sri Lanka.
Ultimately, Dr. de Silva underscored that Sri Lanka must evolve beyond remaining a mere consumer of foreign-developed platforms. By focusing on practical, cost-effective deployments tailored to the country’s budgetary parameters, linguistic diversity, and infrastructural realities, Sri Lanka aims to foster homegrown technological value and deliver tangible benefits across key sectors.