Large Language Models for Small Language Communities

This opinion article argues that developing economies with minority languages are largely excluded from the benefits of large language models (LLMs), which are predominantly trained on English datasets. The author, Vivek Badrinath, proposes that policymakers should partner with mobile network operators (MNOs) to build scalable and inclusive LLMs for these communities. MNOs possess extensive developer networks and data-processing capabilities that can be leveraged to create localized AI tools. The piece highlights the example of Nepal, where mobile technology is increasingly central to daily life, suggesting that similar approaches could empower other small language communities. The article is a forward-looking analysis published on July 22, 2026, emphasizing the need for inclusive AI development to bridge the digital divide.