Embodied interactions

Embodied Interaction refers to the way AI systems engage with the physical world through a body—whether robotic, virtual, or sensor-enabled. Unlike purely digital AI models, embodied AI can perceive, move, and act within its environment, allowing it to learn through direct experience. This interaction grounds the AI’s understanding in real-world contexts, enabling more intuitive and adaptive behaviour.

For example, a robot learning to navigate a room or manipulate objects develops a richer, more practical understanding of space, cause and effect, and human cues.

This concept is central to advancing AGI, as it mirrors how humans learn—not just through abstract reasoning, but through sensorimotor experience and physical engagement with the world.