What is Physical AI?
Also called: Embodied AI
Physical AI controls machines in the real world: robots, autonomous vehicles or drones that perceive with sensors, plan and act.
Example
A warehouse robot uses cameras to recognise parcels of different shapes, picks them up and places them on the right conveyor belt.
For pros
Such systems combine perception (camera, lidar), a model of their surroundings and motion control. Because training with real robots is slow and expensive, they often practise first in physically accurate simulations with synthetic data and are then fine-tuned in the real world. Mistakes here have physical consequences. Testing, emergency stops and human oversight are therefore especially important.
Related prompts, skills, agents and tools
Written by AI Compare Ninja with AI assistance and edited by us. Every explanation is our own text.