Let me explain what an AI robot is. The term ‘robot’ comes from the Czech word ‘robota’, which means forced labor. In 1921, Czech playwright Corel Kapik introduced a three-act play called ‘Rossum’s Universal Robots’. The play depicted humans inventing robots as laborers to be enslaved. However, the robots rebelled against their human creators and declared war on humanity. Eventually, the robots could reproduce and replace the human race. This play was the first to introduce the concept of robots to the public, and the term ‘robot’ became synonymous with mechanical beings.
Traditional robots without artificial intelligence can only perform a series of repetitive actions through programming. Nowadays, they are widely used in various industrial manufacturing settings, such as robotic arms on car production lines.
When a robot is equipped with perception devices to gather information from the environment and autonomously interact with the world, it becomes an AI robot or intelligent robot. One such example is Tesla’s Optimus, which is currently under development.
The industry has a professional term for AI robots called ’embodied intelligence’, which refers to artificial intelligence with a physical body combining AI algorithms in the virtual world with hardware in the physical world.
An intelligent robot consists of five main components:
-
Perception system: These are the sensors that allow the robot to gather information from its internal and external environment, such as radar, cameras, and microphones.
-
Drive system: This system controls the robot’s components to perform various actions and movements, often achieved through motors equivalent to human muscles.
-
End effector system: It is the system the robot uses to physically interact with the outside world, such as robotic hands, grippers, mobile chassis, or legs equivalent to human hands and feet.
-
Power system: This provides energy to the robot, either through a power supply or batteries.
-
Computational system: This is the software that allows the robot to execute commands or autonomously respond to stimuli, running on the computational control board equivalent to the human brain. The robot’s computational system processes information obtained from the perception system and other components, utilizes AI models to make decisions, and then sends instructions to the drive system to precisely control the robot’s movements.
However, there are currently very few AI robots in real-life production scenarios that can execute complex tasks like humans. Household vacuum robots and restaurant delivery robots, for example, have only limited perception abilities, allowing them to perform simple tasks like mapping and obstacle avoidance. The reason is that each of the five components of a robot presents technical challenges, such as drive precision, computation accuracy, environmental perception complexity, and limited data for robot learning, among many others.
You may wonder whether using chatGPT’s capabilities could achieve most of an AI robot’s performance. The answer is no. Human interaction relies heavily on sensory information gathered through vision, hearing, touch, etc. Semantic communication, which chatGPT is good at, only accounts for about seven percent of information exchanged during interactions.
Some have envisioned a future where robot hardware partly uses biological materials, where electrical signals become bioelectrical signals, potentially leading to significant energy savings. Furthermore, binary signals could be replaced by other communication means, similar to quantum superposition, greatly enhancing computational power.
If Optimus Prime were to appear in all aspects of human life one day, would the events of the three-act play repeat themselves? Well, opinions may vary.
