AI tool reconstructs viewed images from brain scans with high precision

This digest was compiled by AI from multiple sources — links to the originals are below.
Researchers at the Weizmann Institute of Science developed an AI tool that reconstructs images a person views from fMRI brain scans with remarkable precision. The tool also predicts brain activity from images, potentially aiding communication for locked-in patients. Ethicists caution the technology could extract private thoughts without consent.
Key Facts
- The AI tool was developed by Michal Irani and colleagues at the Weizmann Institute of Science in Rehovot, Israel.
- The tool uses higher-resolution fMRI datasets where each voxel covers about one cubic millimeter, compared to typical scanners' three cubic millimeters.
- Judy Illes, a neuroethicist at the University of British Columbia, described the work as 'magnificent' and therapeutically exciting.
- Tommy Sprague, a neuroscientist at the University of California Santa Barbara, warned that surreptitious extraction of thoughts could make '150 years of sci-fi come true'.
Mind-Reading Technology
The AI tool guesses what a person is looking at by analyzing brain scans and recreates the image with remarkable precision. It can also predict a person's brain activity based on what they are viewing. Michal Irani, who developed the tool with her colleagues at the Weizmann Institute of Science in Rehovot, Israel, hopes it will reveal more about how the brain works. Potential applications include helping locked-in people communicate and allowing scientists to recreate the content of dreams.
Ethical Concerns
Judy Illes, a neuroethicist and professor of neurology at the University of British Columbia, described the work as 'magnificent' and the therapeutic potential as 'tremendously exciting'. Tommy Sprague, a neuroscientist at the University of California Santa Barbara, said the results seem very impressive but warned that surreptitious extraction of thoughts could make '150 years of sci-fi come true'. Other scientists warn that a similar approach could reveal inner thoughts and mental imagery potentially without consent.
Technical Advances
Neuroscientists have worked for years on reconstructing what people see, with early attempts producing blurry images. Irani and her colleagues used newer datasets from higher-resolution fMRI scanners, where each voxel covers around one cubic millimeter of neurons instead of the typical three cubic millimeters. Other teams have also developed tools to recreate images from brain scan data, but Irani says they are not good enough.