Google's AI Research Assistants Now Used by 100% of Team, Evaluating 200,000 Models at Once to Accelerate Scientific Discovery

Google Research's AI-powered Empirical Research Assistants are now used by 100% of its team, evaluating up to 200,000 computational models simultaneously to accelerate breakthroughs in genomics, neuroscience, and epidemiology, while also tackling science's reproducibility crisis — though human expertise remains critical to prevent AI from generating physically impossible results.

• Google Research's AI-powered Empirical Research Assistants (ERA) are now used by 100% of the team, enabling scientists to evaluate up to 200,000 computational models at once instead of just a handful, dramatically accelerating discovery across genomics, neuroscience, and epidemiology.
• AI agents are tackling a long-standing reproducibility crisis in science by re-implementing research methods from just a few sentences of description and generating comparable results, with the technology being rolled out broadly through the Gemini for Science product line.
• Human expertise remains essential, as AI systems can produce physically impossible solutions without rigorous expert verification, and the biggest scientific breakthroughs continue to emerge from systematically solving smaller bottleneck problems rather than single dramatic leaps.