Understanding the concepts behind the machines with human-like capabilities.
What is AI?
Artificial Intelligence (AI) refers to the ability of machines to replicate an approximation of human intelligence. The term is used to describe applications that can reason, discover, learn, generalize, plan, problem-solve, make decisions, recognize voices and images, understand languages, perform tasks that require complex thinking, accuracy and speed, as well as other human-like capabilities.
Levels of AI
There are four levels to AI which describe their degree of sophistication and human-like capabilities.
Reactive AI is the most basic and produces outputs based on input with no chance of variation. It responds to identical situations in an identical fashion with no learning ability. One notable example of this is IBM Chess program, Deep Blue, which beat world champion Garry Kasparov.
Limited memory AI is the most used kind of AI today and was developed from the foundations of reactive AI. It uses pre-programmed information to react to observational data and is somewhat able to learn but has no long-term memory. An example of this is self-driving cars.
Theory of mind AI goes a step further with a focus on emotional intelligence. It is able to consider subjective elements such as emotions then adjust accordingly when making decisions. Although there are no true examples, the humanoid robot Sophia by Hanson Robotics in Hong Kong debuted in 2016 has shown a glimpse into what a theory of mind AI robot might be like.
Self-aware AI is the most advanced level and refers to a machine with an equivalent consciousness and intelligence to a human being. It is capable of thinking for itself, experiencing emotions, and acting independently. The self-aware AI is the level of AI often depicted in movies featuring AI such as Skynet in Terminator and Ultron in Avengers: Age of Ultron.
English 
日本語