New insights from Google’s AI & Economy ATLAS
Google's AI & Economy ATLAS is getting a new interactive, open-access experience that makes its millions of global data points easier to explore. Users can now examine how different occupations — from electricians to purchasing managers — are using AI, look at how people use AI at home, and compare AI adoption rates across countries.
The ATLAS occupational data shows distinct regional patterns. India's creative industry uses AI at a higher rate than the rest of the world, with arts, design, and media occupations making up 19% of work-related AI usage — 1.6 times the global average. The U.S. leads in technical AI adoption, with computer and mathematical occupations accounting for 30% of work-related AI usage, double the share elsewhere. In OECD countries, computer/mathematical and business/financial operations occupations lead AI usage, while in non-OECD countries the leaders are office and administrative support, arts/design/entertainment/sports/media, and educational instruction and library roles. Adoption broadly correlates with national income, but Brazil and the UAE stand out with higher adoption than their GDP per capita would predict. Usage for real-time equipment diagnostics and troubleshooting also varies: Brazil and Germany direct 7% of work AI usage toward manual tasks (1.4 times the global average), versus 4% in Japan.
New ATLAS-based research from Google and Google DeepMind, in collaboration with MIT FutureTech, examines how scientists use AI. Nearly half of surveyed scientists use some form of AI every day — a higher rate than many other occupations — and both LLMs and specialized models are used widely for science in mutually reinforcing ways. Scientists report saving almost 7 hours a week with AI, freeing time for more research, but that gain has created bottlenecks further down the research production pipeline, leaving a backlog of hypotheses.
The combined release points to AI reshaping global work unevenly across occupations and regions while delivering measurable time savings in science that are not yet translating into faster end-to-end research output.