Embodied AI integrates AI into physical bodies such as robots to enable them to perceive, learn and dynamically interact with the environment like humans. Relying on computation and logical reasoning, it can continuously learn, adapt and complete tasks by perception, action and environmental feedback, thus improving autonomy and practical application in real world.

In addition to humanoid robots, other hardware with interactive capabilities, such as autonomous driving vehicles and wearable devices, can also serve as bodies of embodied AI.

Many cities in China are now funding projects that advance embodied intelligence systems, allowing AI to physically interact with the world.