Google Study: AI Use Spans 70% of Jobs But Automates Few Tasks

Artificial intelligence tools have reached workers across a remarkably broad range of occupations, but the technology is functioning primarily as a collaborative assistant rather than the job-eliminating force many have feared, according to new research from Google.

The company's ATLAS study — Activity, Task, Landscape and Adoption Study — examined 14.65 million de-identified interactions with its Gemini AI across the consumer app, AI Mode search, and API over a two-week period in April. The findings reveal a significant gap between the theoretical scope of AI's capabilities and how workers actually deploy the technology day-to-day.

Why it matters

While speculation about AI-driven job displacement dominates headlines, this large-scale analysis of real-world usage shows workers are integrating AI into existing workflows rather than handing over entire roles. The data suggests organizations may need to rethink both their automation expectations and their AI training strategies to reflect how employees actually use these tools.

Breadth Without Depth

Scott Strand, an economist on Google's Technology and Society team, characterized the pattern as "very, very broad" adoption touching a huge range of occupations. According to the research, Gemini usage appeared across 70% of job categories, representing approximately 90% of U.S. employment.

However, Strand noted that AI use remains "very shallow." The average worker uses AI for only 21% of their tasks, and fewer than 10% of Gemini interactions appeared designed to automate non-routine cognitive work. Instead, workers primarily use the tool for research, drafting, iteration, troubleshooting, and learning — collaborative functions that augment rather than replace human judgment.

Methodology and Limitations

Google's automated systems classified interactions as work-related or personal, then mapped work activities to 4,000 tasks across 800 occupations in 150 countries and 140 languages. Non-work tasks were categorized using Bureau of Labor Statistics frameworks.

The study excluded data from business-focused products like Gemini Enterprise and Google Workspace because the company doesn't maintain logs for those services. This gap means the research may understate automation levels in enterprise environments where AI adoption could be more advanced.

Unexpected Patterns

The research uncovered several surprising trends. Blue-collar workers, including electricians seeking wiring diagrams and auto repair technicians looking for engine maps, used AI more than researchers anticipated. These workers frequently employed multimodal AI capabilities involving images and video.

The study also found that higher-paid workers and those in wealthier regions use AI more intensively. A 1% increase in an occupation's median earnings correlated with a 2.68% increase in AI usage levels.

Consumer applications dominated overall usage, accounting for 86% of Gemini conversations. Beyond completing household tasks more efficiently, people used the AI to navigate government services and processes.

Looking Ahead

Whether studies like ATLAS can predict future AI adoption patterns or merely document current usage remains an open question. Google plans to publish additional reports in the series, though the company hasn't determined the update frequency.

The findings align directionally with earlier research from Anthropic and OpenAI, though Anthropic reported somewhat higher automation levels. Details of the Google study were first reported by Automation Watch.