The Branches of Artificial Intelligence Explained

·4 mins
Christoph C. Cemper

Artificial intelligence is a field of computer science centered around creating machines that can think like humans can. In theory, a flawless AI would be able to learn from past experiences and use advanced reasoning to perform tasks faster and more accurately than humans can. Different forms of AI are created using advanced algorithms that allow them to process data and make decisions based on their understanding. The capabilities of these types of AI vary, and the limits of what AI can do are constantly expanding as research and development continue.

Artificial Narrow Intelligence (ANI)>

Artificial Narrow Intelligence (ANI) #

Artificial Narrow Intelligence (ANI) describes an AI that is designed to carry out specific commands and actions. An ANI is built to excel at one type of cognitive ability and can’t learn new skills independently. Siri on Apple iPhones is an example of a narrow AI; it has a limited, defined range of functions.

Artificial General Intelligence (AGI)>

Artificial General Intelligence (AGI) #

Artificial general intelligence does not yet exist, but in theory, it’s AI that can learn and think just like a human can. AGI can apply knowledge and skills from different contexts to a problem or situation. Also called “strong AI,” AGI can learn independently from its environment. For now, true AGI only exists in science fiction, though research is ongoing to make it a reality.

Artificial Super Intelligence (ASI)>

Artificial Super Intelligence (ASI) #

Artificial super intelligence (ASI) is a type of AI that has the capability to do tasks beyond what the human mind can think or comprehend. This AI is capable of manifesting cognitive skills and developing thinking skills of its own. An artificial superintelligence would be completely self-aware and able to make decisions and judgments on its own.

Reactive Machines>

Reactive Machines #

A reactive machine is an AI system that is made to do a specific task and does not have the ability to learn. These artificially intelligent machines react depending on the input given and only work with the present data. One of the most famous examples of a reactive machine is Deep Blue, the IBM chess-playing AI. It could analyze a chess board to predict future moves and select the best move to make, but it could not learn from its mistakes to improve its skills.

Limited Memory>

Limited Memory #

As the name suggests, limited memory AI can remember and learn from past experiences, but these experiences are only held temporarily in its memory. Limited memory is crucial to the process of machine learning; the AI needs to be able to remember information in order to extract knowledge from it. ChatGPT is a popular example of limited memory AI; it’s been trained to know a lot of things, but as you work with it, you’ll notice that it can only remember the information in the  prompt you give it for a short time before it starts to forget.

Theory of Mind>

Theory of Mind #

Theory of mind AI can have human-like responses to its environment, including understanding emotional cues. It can demonstrate knowledge of how humans think and behave and adjust its behavior accordingly.


Self-Awareness #

Self-aware AI currently only exists hypothetically. A self-aware artificial intelligence would have full consciousness; it would be able to have its own feelings and desires as well as understand the emotions and needs of the humans around it. Self-aware AI does not exist; it’s the most  advanced type of AI and the ultimate goal of many AI researchers.