Scientists save nearly seven hours a week with AI, but verification backlogs grow

Scientists using AI save nearly seven hours per week, with the freed time mostly reinvested into more research — that is the figure from Google's research.

Before this, no such cross-stage quantification existed: AI had entered daily research work, but the scale of time savings and their costs lacked systematic measurement.

The study by Mihai Codreanu and colleagues draws on three data sources: a sample of 15 million Gemini interactions, an inventory of over 2,600 discipline-specific AI models, and a survey of more than 600 scientists. Nearly half of surveyed scientists report using some form of AI every day. As analysis and writing get faster, bottlenecks move downstream — scientists report a growing backlog of untested hypotheses and rising demand for output verification. Note this is Google's own research, with usage proxied through Gemini and the time savings self-reported by scientists.