Building AI for All: Language, Edge and the Future of AI

Artificial intelligence is reshaping economies globally and, as a general-purpose technology, its adoption has been faster than any previous digital innovation. In less than three years, more than 1.2 billion people have used AI-enabled tools, outpacing the early adoption of both the internet and smartphones. Yet AI adoption remains deeply unequal, with usage rates in high-income economies roughly twice those observed in low- and middle-income countries (LMCIs). 

Most AI systems are designed for environments with abundant compute infrastructure, high-speed connectivity and massive datasets – conditions that remain limited across many emerging markets. As a result, many AI tools remain inaccessible, unreliable or irrelevant for large segments of the global population, reinforcing existing digital divides. 

In this context, a growing number of innovators and mobile operators are exploring an alternative approach: designing systems that adapt to local realities, particularly in LMICs. This includes models that work in local languages, building practical data pipelines and deploying AI at “the Edge” through decentralised infrastructure. 

Drawing on new GSMA research and real-world innovations from startups deploying AI across various sectors, this session will explore the potential of local language AI and edge computing for resource-constrained environments and why data quality, accessibility and representation in LMICs are vital to delivering equitable, context-appropriate solutions. It will examine how adapting models and infrastructure can bring AI closer to users and how mobile operators, innovators and ecosystem partners can anchor a more inclusive AI paradigm.