Introduction

The emergence of large language models (LLMs) has kindled discussions about the potential for artificial intelligence (AI) to increase productivity and boost economic growth. Although the economic effects of AI have been a topic of debate for the last decade, the growing adoption of LLMs creates a new urgency for understanding the technology’s impact.

To date, most research on the economic effects of AI has centered almost entirely on the Global North, yet the diffusion of LLMs around the world economy may have significant implications—both auspicious and foreboding—for the Global South. On one hand, the productivity-enhancing benefits of LLMs could greatly benefit industries in the Global South. On the other hand, uneven adoption of LLMs could exacerbate the digital divide and lead to an uneven distribution of the technology’s benefits.

Preliminary evidence suggests that LLMs are improving productivity across a variety of white-collar tasks, including writing, coding, and customer service. Such gains have been particularly concentrated in the IT sector. IT services have played a growing role in the developing economies of the Global South, many of which export key digital services to the wealthier Global North. This includes telecommunication services, copywriting, gig work, and other forms of content generation. The efficiency gains ushered in by LLMs may be disruptive to the growth of these IT sectors, reshaping the balance and flow of international IT exports.

This report investigates the potential impact of LLMs on the Global South’s IT sector. It begins by contextualizing the growing role of IT in major Global South economies, and then highlights the kinds of IT occupations that are most represented in Global South countries. It concludes with a discussion regarding the impact of LLMs on the global flow of IT services and an assessment of current policy responses to AI by Global South countries.

This report finds that growth in the IT sector’s share of Global South employment and exports is likely to be disrupted due to the infusion of LLMs across the world economy. Although Global South countries experience comparatively low levels of LLM exposure due to smaller shares of IT services, the IT services that are most represented in Global South exports, growth, and employment tend to be ones with high levels of automation potential. In other words, the IT services in the Global South appear more likely to experience the displacing, as opposed to complementing, effects of AI. Given the potential for countries to reshore and automate previously outsourced IT occupations, the Global South’s IT services appear vulnerable to LLM adoption. Many countries are aware of these risks and are moving rapidly to promote reskilling and develop diversified and more advanced IT sectors. Yet, existing policy responses may be insufficient to address these risks in the Global South.

To address these challenges, policymakers should take three key steps:

1. Policymakers in the Global South should support workforce development policies that provide workers with the new digital skills they will need for the AI economy. AI will reduce demand for some digital services but increase demand for others, and economies in the Global South will need to adjust.

2. Policymakers in these countries should pursue widespread adoption of AI to boost productivity and competitiveness across their economies and develop domestic AI implementation skills and capabilities.

3. Policymakers should continue to pursue policies that facilitate digital free trade, such as by opposing restrictions on cross-border data flows, to ensure that their companies and workers have access to best-in-class digital services.

Read the full report. (PDF)