Google Introduces ERA, an AI Tool for Scientific Coding Described in Nature

Google Research announced Empirical Research Assistance (ERA), an AI tool for expert-level scientific coding, described in a paper published in Nature. ERA uses Gemini to write and optimize scientific code, addressing the iterative testing and refinement of computational experiments. Given a scientific problem and a success metric, ERA searches literature, writes code, explores solutions, and evaluates results using tree search. Tested on benchmarks in genomics, public health, satellite imagery, neuroscience, time-series forecasting, and mathematics, ERA achieved expert-level performance across all. ERA is one of the systems used to build Computational Discovery, a new experimental tool rolling out through Gemini for Science. Google also launched Hypothesis Generation, built with AI Co-Scientist, also published in Nature today.