Global South trails in generative AI, Microsoft finds

Businesses in the Global South use generative AI at lower rates than peers in wealthier countries, industry figures and regional experts say. Microsoft’s 2025 data shows generative AI diffusion at 14.1% in the Global South and 24.7% in the Global North. The company measures diffusion as the share of people who used a generative AI tool during the second half of the year.

The gap is not uniform. Several countries in Asia and the Gulf report high use, with two economies registering diffusion rates well above many Western countries. Experts say those outliers underscore differences between and within regions often grouped together as the Global South.

Industry leaders and researchers identify three main barriers to wider adoption: limited local data, weak computing and energy infrastructure, and migration of skilled workers. Arun Kumar, global head of product and industry practice at Altimetrik, described talent loss as a major challenge: “Brain drain for us is a true challenge.” He pointed to language and data quality as related problems, noting that many early AI systems were trained on English-language data and can miss local languages and conversational patterns.

Kumar used India as an example, where many languages and dialects are spoken. He said call-center recordings and other interactions can lose nuance when conversations occur in local languages but are recorded or summarized in English. Kumar pointed to government-supported language projects designed to add local linguistic expertise into models.

Infrastructure limits include restricted access to advanced processors, fewer local data centers and unreliable electricity. Those constraints make it harder for companies to build and run large models on-site, increasing reliance on foreign cloud and compute services. Observers also note a concentration of compute and investment in wealthier markets, and officials held an international summit in the Global South in February to discuss broader access and governance.

Some researchers warn against treating the Global South as a single bloc. A number of analysts say the dominant AI approaches developed in major economies do not align with the priorities or needs of many countries outside those markets, forcing policymakers and businesses to weigh different models of development and regulation.

Small-business operators report uneven awareness and use of AI tools. Jason Mann, founder of a marketing agency that hires remote staff in the Philippines, trains employees to use tools for research and reporting and said AI is common workplace talk in Western markets but less widespread among some Global South teams. He highlighted attributes he looks for when hiring: work ethic and curiosity. “Having a curious mind is the most important skill that you can have,” Mann said.

Microsoft’s diffusion figures and the accounts of practitioners and researchers detail where adoption is concentrated and where gaps remain. Data quality in local languages, available compute and trained personnel shape how quickly companies can deploy generative AI tools.