AI Scientist Autonomously Generates and Validates Biological Discoveries
Researchers have developed an AI scientist capable of generating hypotheses, designing experiments, and interpreting results, advancing self-driving laboratories with minimal human intervention.
Researchers at Chalmers University of Technology in Sweden have created an AI scientist that can autonomously generate scientific hypotheses, design experiments, and interpret their results. This represents a significant advance in self-driving laboratories, reducing the need for direct human involvement.
The AI system combines advancements in large language models, automated reasoning, and laboratory automation. The researchers built a closed-loop AI laboratory that has been used to conduct research on the brewer's yeast, Saccharomyces Cerevisiae. The AI was initially provided with data on the yeast's basic metabolism as its research foundation.
During the process, the AI formulated hypotheses regarding how specific genes could be modified to improve yeast growth. It then designed experiments to test these hypotheses and utilized automated laboratory systems to execute them. Finally, the AI analyzed the experimental outcomes and generated new hypotheses for subsequent investigation.
According to the researchers, this capability to independently generate and validate new discoveries is a crucial step towards fully autonomous scientific laboratories. The aim is to accelerate scientific discovery and enable a deeper understanding of complex biological systems.