tech
Google study finds broad AI use, but little evidence of job automation
An analysis of nearly 15 million Gemini interactions found people are mostly using AI to collaborate, research and learn
TL;DR
- AI use is broad, touching 70% of jobs which represent 90% of U.S. employment, according to a Google study of Gemini interactions.
- Workers primarily use AI as a collaborator for research, drafting, iteration, troubleshooting, and learning, not for delegating entire jobs.
- AI use is shallow, with the average worker using it for only 21% of tasks, and less than 10% of interactions aimed at automating non-routine cognitive work.
- The study analyzed 14.65 million de-identified interactions with Gemini across its app, AI Mode search, and API over two weeks in April.
- Workers in higher-paid jobs and wealthier geographies tend to use AI more.
- Blue-collar workers are using AI, particularly multimodal AI involving images and video, for tasks like seeking diagrams and maps.
- Non-work consumer use accounted for 86% of Gemini conversations, with users employing it for everyday tasks, government services, and processes.
- The study did not include data from business-focused tools like Gemini Enterprise and Google Workspace, which could show higher automation levels.