The United States is actively developing AI systems designed to train robots for real-world tasks, with companies employing various methods to achieve this goal. One approach involves teleoperation, where humans use controllers to remotely guide robots through tasks, enabling the AI to learn from successful and unsuccessful attempts. Another method includes capturing human movement data, where individuals are compensated to wear smart glasses that record their daily actions.
Companies like Tesla, Boston Dynamics, and Nvidia are leading this technological advancement, with Nvidia projecting a $38 billion market for humanoid robots within the next decade. Evercoast, an AI software company, is collaborating with the U.S. Air Force to train robots in repairing fighter jets. They utilise human demonstrations to create simulations, which then guide robots in performing these tasks.
This focus on physical AI is creating a new industry centred around human movement data collection, potentially leading to the next generation of autonomous robots in homes and workplaces. The ability to rapidly train and retrain robots is seen as a crucial factor in determining future manufacturing leadership.
