A Bernstein study tracking global AI adoption through March 2026 concluded that emerging markets are deriving outsized value from artificial intelligence relative to developed economies.
The research found that nearly 18% of the global working-age population was using AI by March 2026, up from 15% nine months earlier. That rise in absolute users underpins Bernstein's central challenge to the commonly held view that advanced economies are markedly ahead in practical AI uptake.
Bernstein cautioned that simple per-capita comparisons are often misleading. Emerging markets typically have larger populations and a higher share of workers in agriculture and low-end manufacturing, sectors in which AI adoption is lower. The firm said that when AI engagement is measured in absolute terms within the information economy - the subset of the workforce most likely to use AI tools - the apparent advantage of developed markets diminishes substantially.
The report also identified regional differences in how AI is applied. In emerging markets, AI activity is concentrated in software development, writing, editing and education. Developed markets, by contrast, show greater AI usage in sales, finance and healthcare. Bernstein described middle-income economies as having a higher concentration of AI use in a few areas, whereas high-income economies display broader adoption across many sectors.
The study quantified productivity effects and capability gains. Users in emerging markets reported average time savings of 4.6 hours per task, compared with 3.8 hours per task in developed markets. Bernstein also found that 16% of tasks delegated to AI in emerging markets were beyond what users could have performed themselves, compared with 12% in high-income economies.
Bernstein noted a difference in intent: developed markets tend to use AI primarily to enhance quality, while emerging markets devote more than half of their AI activity to automation. The firm additionally observed an 8% increase in weekend AI activity among top-skilled workers, with software engineers shifting some of their effort from weekday debugging to weekend experimentation.
The study's findings emphasize variation in adoption metrics, usage patterns and outcomes across income groups and regions, while highlighting the notable role of automation and productivity improvements in emerging economies.
Clear summary
Bernstein's March 2026 analysis shows nearly 18% of the global working-age population using AI, a rise from 15% nine months earlier. The firm argues that absolute usage within the information economy narrows the adoption gap between developed and emerging markets. Emerging markets concentrate AI in software, writing, editing and education, report larger per-task time savings, and allocate a majority of their AI activity to automation. Developed markets emphasize quality improvements and broader sectoral spread.
Key points
- Measured by absolute usage in the information economy, the developed-emerging adoption gap is much smaller than per-capita comparisons suggest - impact on tech and services sectors.
- Emerging markets focus AI on software development, writing/editing and education while developed markets lead in sales, finance and healthcare - implications for software, finance and healthcare sectors.
- Users in emerging markets report larger time savings per task (4.6 hours) and a higher share of tasks (16%) performed by AI that users could not do themselves, versus 3.8 hours and 12% in high-income economies - productivity and labor implications for information work.
Risks and uncertainties
- Per-capita metrics can misrepresent comparative adoption because of demographic and sectoral differences, creating the risk of misleading conclusions for policymakers and investors in labor-intensive industries.
- Concentration of AI use in a few areas in middle-income economies may raise exposure to sector-specific shocks and limit broader economic diffusion, affecting sectors such as education and software services.
- Differences in the purpose of AI - quality enhancement in developed markets versus automation in emerging markets - introduce uncertainty about longer-term impacts on employment patterns in affected sectors.