Are brain waves the next unlock for physical AI?

A company building data tools for AI training is experimenting with brain wave sensors to improve the training data used for robots and humanoid machines. The brain wave measurements are meant to capture mental states like error, intent, and surprise, with the theory that this additional information could help AI models learn physical tasks more effectively. The core challenge these companies face is that training AI for physical manipulation requires vastly more real-world data than text-based AI needed, and this data must be painstakingly manufactured through human demonstrations rather than scraped from the internet. This data creation has become its own business because the sheer volume needed to teach robots complex physical skills does not exist readily available.