Researchers at Stanford School of Medicine have unveiled an innovative AI-driven virtual lab aimed at accelerating scientific discovery, featuring an AI principal investigator and expert scientists.

This Virtual Lab operates similarly to traditional research environments, with the AI PI coordinating expertise and agents to address scientific challenges, including the development of vaccines for SARS-CoV-2.

In a recent task, the AI scientists proposed using nanobodies—smaller and simpler antibody fragments—to enhance the SARS-CoV-2 vaccine, resulting in a solution with greater stability and binding affinity to COVID-19 variants.

The AI's choice of nanobodies was driven by their computational advantages and unique properties in binding to viruses, showcasing the AI's decision-making capabilities.

The Virtual Lab team successfully designed 92 new nanobodies, with two demonstrating promising results against SARS-CoV-2 variants in laboratory tests.

The virtual lab operates efficiently, conducting rapid meetings and discussions, capturing all interactions for human researchers to monitor progress and enhance transparency.

This collaboration between AI agents and human researchers has significantly increased the number of testable hypotheses, thereby enhancing overall research capacity.

James Zou, PhD, who led the project, emphasized the importance of interdisciplinary collaboration and the capabilities of AI systems in tackling complex biological problems.

The study highlights the potential of AI not only as a tool but as a primary driver in scientific discovery, indicating a new paradigm in research methodologies.

The findings suggest that the Virtual Lab could facilitate rapid and impactful scientific discoveries, particularly in the field of virology.

Zou's team is enthusiastic about applying the virtual lab's capabilities to a broader range of scientific inquiries, including reassessing published research for new insights.

The study, published on July 29, 2025, in Nature, illustrates how these AI agents can effectively design nanobodies against new variants of the SARS-CoV-2 virus.