Thousands of AIs Analyze Millions of Brain Cell Data in Just 32 Minutes, Uncovering New Clues for Alzheimer's and Parkinson's Research

What happened

OpenAI announced on October 21, 2026 that it used thousands of AI agents to study experiment data from millions of mouse brain cells. The analysis took only 32 minutes. Human researchers would likely need months to a year for the same job. The AI also suggested new ideas about how some genes work, ideas that people may have missed.

Key numbers and facts

• Time taken: 32 minutes
• Data: millions of mouse brain cells
• Number of AI agents: thousands
• Human estimate: several months to a year
• One AI agent alone cited 11 papers in the largest test. Thousands of agents cited thousands of papers.

Background

Scientists can now test how a specific gene affects cells. They use a tool called CRISPR, which changes particular genes. The team used a dataset called the whole-brain CRISPR atlas, which records how mouse brain cells react when genes are changed. Finding useful clues in such large data takes a long time, and small changes can be missed. The AI agents split up the work. Each one studied a different gene, checked cell changes, searched science papers, and suggested reasons for what it saw. Then the results were combined into a map called an 'inference atlas.'

Two example findings

The AI found that Tsc1 and Tsc2, two genes linked to a system called mTOR that controls cell growth, behave differently in the cerebellum, the part of the brain that helps control movement. Changing Tsc2 changed dozens of genes that help nerve cells talk to each other. Changing Tsc1 changed very little. Tsc2 problems are linked to tuberous sclerosis, a disease that causes benign (not cancerous) tumors. The AI also looked at DCTN1, a gene linked to Perry syndrome, a disease with Parkinson-like symptoms. Changing Dctn1 raised activity in genes that make cholesterol. The AI suggested that if the system that moves cholesterol inside nerve cells breaks, cells may make too much of it.

Why it matters

This method could help scientists sort through huge amounts of data faster. It can point them to the cell types and genes most worth testing. The team says it could help check whether existing medicines might help. The researchers still need to test these ideas in the lab. OpenAI says scientists will still lead the research, and that AI can help find better ideas to test, which may lead to better medicines in the future.