Resources / Recordings / Brain Tuning: Using Human Brain Data to Improve AI

Recording

Brain Tuning: Using Human Brain Data to Improve AI

With David Markowitz


Date

Human brain recordings are becoming a credible resource for improving AI. Early studies show that neural data can improve language and reasoning models, without first needing a complete explanation of how the brain computes. This talk makes the case for “brain tuning”: using human neural activity as an additional training signal to build more capable – and potentially more efficient – AI. I will examine what the evidence establishes, why human cognition is a promising target, and what it would take to extend these results to frontier models, starting with which recordings are worth collecting at scale. The implications extend beyond AI: if brain tuning delivers gains that justify continued collection, AI demand could finance neuroscience infrastructure at scale, driving a flywheel that accelerates brain-computer interfaces, connectomics, and eventually whole-brain emulation.

Fund the science of the future.

Donate today