AI Development Widens Global Digital and Economic Divides

Artificial intelligence is rapidly integrating into various sectors, from email assistance and job application filtering to education and healthcare. Despite industry leaders promoting "AI for everyone" and governments developing national AI strategies, a decade of work in digital inclusion across Europe, sub-Saharan Africa, and Southeast Asia reveals a recurring pattern: new technologies exacerbate existing inequalities in connectivity, skills, and institutional capacity. The current AI wave amplifies these fractures, creating a divide not just in technological access but also in economic opportunity and global influence.

Compute power for AI is heavily concentrated, with the United States hosting over 5,000 data centers, significantly more than any other nation. This concentration, coupled with the reliance on cloud platforms, means AI development is largely outsourced, controlled by a few states and firms. Systems trained in narrow institutional and linguistic environments may struggle to serve a global population effectively. Furthermore, AI skills and literacy are deeply stratified. While advanced AI competences are concentrated among highly educated workers, basic digital skills remain limited for many adults. Governments are attempting to integrate AI into education, but support for AI literacy at lower educational levels and for educators is inconsistent. Individuals with strong educational backgrounds and digital skills are better positioned to leverage AI, while others risk experiencing it as an opaque system that impacts them without their input.

Agenda-setting power in AI largely rests with a few industry actors and technologically advanced states, leaving most countries in a reactive position. Communities generating vast amounts of data often have little say in AI system design or governance, with their languages and experiences underrepresented. Efforts in countries like South Africa and Indonesia highlight both challenges and possibilities. South Africa's attempt to release a draft AI policy was marred by AI-generated citations, exposing a gap between governance ambitions and institutional capacity. Indonesia, through its National Research and Innovation Agency (BRIN), is developing practical AI tools for underserved communities, including applications for artisanal fishermen and multilingual language models. Regional cooperation, such as Africa's Continental AI Strategy, aims to reduce dependence on externally developed AI systems by fostering local-language models and open-source ecosystems. The AI divide extends beyond access to AI itself, encompassing disparities in who can shape its purpose, the problems it addresses, and the societal priorities it serves.