Since the public emergence of ChatGPT in 2023, the use of large language models (LLMs) has spread unevenly across regions, with professionals in Global South nations adopting more ambitious applications to fill critical gaps in law, education, and journalism, while many in the Global North take a more cautious approach.
- Global South practitioners use LLMs for high-impact, resource-filling roles.
- Global North experts generally exhibit more risk-averse LLM usage.
- Access barriers and accuracy concerns temper the enthusiasm for LLMs.
What happened
Following the public launch of ChatGPT in early 2023, diverse professional fields—particularly law, education, and journalism—began experimenting with large language models (LLMs). To understand these evolving applications, researchers conducted interviews with 56 experts from 24 countries spanning seven world regions, capturing a wide range of LLM use cases and perceptions.
The study revealed significant geographic variation in LLM adoption. Practitioners in the Global South frequently deploy LLMs to perform complex, high-value tasks, often stepping in to fill critical service gaps where legal, educational, or journalistic capacity is limited. Conversely, professionals in the Global North generally approach LLMs more cautiously, showing skepticism rooted in concerns about risks and potential harm.
Why it matters
The varied adoption of LLMs highlights the potential for these technologies to support fundamental human rights, such as access to justice and informed journalism, especially in under-resourced areas. For example, in Singapore, LLM-powered chatbots assist litigants in small claims tribunals by providing legal information without incurring lawyer fees, enhancing equity in justice access.
At the same time, uneven access to essential infrastructure like devices and reliable internet connectivity restricts who benefits from these advances. Moreover, reliance on LLMs introduces risks related to accuracy and quality, which can impact legal outcomes, journalistic integrity, and educational effectiveness. These complexities underscore the importance of contextualizing LLM deployment within diverse local realities.
What to watch next
Monitoring how LLMs continue to evolve in both Global South and North contexts will be crucial, particularly in terms of regulatory frameworks, ethical standards, and technical improvements to reduce risks. Insights from practitioners actively using LLMs in resource-limited environments may inform more inclusive, equitable policy approaches addressing the digital divide.
Additionally, stakeholders should watch for innovations that improve multilingual capabilities, accuracy, and user training to maximize the positive impact of LLMs. Ensuring transparency and accountability in automated tools remains vital as these technologies become increasingly integrated into sensitive sectors like justice, education, and news media.