An army of AI agents investigates millions of cells in 32 minutes!
OpenAI, PerturbAI and Scripps Research used thousands of intelligent agents to examine data from millions of mouse brain cells. The analysis, completed in about 32 minutes, generated biological hypotheses to guide further studies. The results show how AI can help organize large volumes of evidence, although the findings require experimental validation.
Biological research generates increasingly large datasets. Interpreting the effects of genetic changes on millions of cells requires linking experimental results with knowledge accumulated in specialized publications. This work can demand a great deal of time from highly trained scientists.
The collaboration used thousands of agents based on GPT-Rosalind to distribute interpretation tasks and explore relationships between results and the scientific literature. The announced analysis was completed in about 32 minutes. The approach does not replace experiments: it helps select the hypotheses that deserve to be studied in greater detail.
This approach looks promising for research on neurodegenerative diseases and other complex conditions. Even so, a relationship identified by an agent does not prove causality, nor is it equivalent to discovering a treatment. The next step falls to researchers who can reproduce and verify the observations.


