MIT researchers develop AI method to translate hand gestures into robot training data

Researchers at MIT have devised an AI approach that captures hand gestures. The captured gestures are transformed into training data for robots. The method uses computer vision to interpret motion patterns. It aims to simplify the generation of robot learning datasets. By leveraging natural human gestures, the technique reduces manual labeling. Early tests show robots can learn tasks from the gesture‑derived data. The work could accelerate development of more intuitive robot interfaces. MIT plans to publish further results and explore commercial applications.