Stanford Graduate School of BusinessAdapt AI to the Local Context: Gaurav Nayyar, World Bank Group
Stanford Graduate School of Business
Adapting AI to local context 0:01
For developing economies, the goal is not to race toward cutting-edge AI but to adopt available tools and adapt them to local languages, challenges, and conditions. Governments can act as enablers by funding electricity and internet infrastructure, as users by piloting AI in agriculture, education, and health, and as regulators by leaning on existing laws and voluntary industry standards rather than writing rules from scratch.
Back office use outpaces the hype 2:32
Despite excitement about AI transforming classrooms and clinics, most current use in developing countries involves government back office functions like predictive analytics, early warning systems, and smart agriculture, relying more on older predictive models than generative AI.
Fast reach, shallow use 4:31
Middle-income countries generated half of ChatGPT's traffic within six months of launch, far faster than steam power, electricity, or the internet once spread, yet only about 5 percent of their internet users use ChatGPT compared with 25 percent in high-income countries.
Jobs face less immediate disruption 7:00
Fewer than 10 percent of jobs in developing economies are currently automatable, since most work is manual rather than cognitive, meaning AI is more likely to complement scarce doctors, agricultural experts, and officials than replace workers, though middle-class knowledge jobs may automate over time.
Analog complements still matter 9:31
Beyond data and connectivity, businesses across seven surveyed developing economies cited lack of knowledge and lack of finance as top barriers to adopting AI, showing that skills, awareness, and access to capital remain essential alongside digital infrastructure.
AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.
