Three Shades of AI: Narrow, General, and Superintelligent

The types of Artificial Intelligence (AI) can be divided into three broad categories - Narrow AI, General AI, and Superintelligent AI.

Narrow AI, or Weak AI, is the most common type and is what’s used for most applications today. It is able to learn and adapt within its own domain, and is used for specific tasks such as facial recognition, speech recognition, natural language processing, and recommendation systems. It typically excels at optimising tasks, and can scale better and faster than humans in many areas. It is mostly useful for automation of simple tasks where the inputs and outputs are well defined, but it is limited by its restricted domain.

General AI, or Strong AI, is the challenge of creating a machine that exhibits both human-level intelligence, or greater. It works across multiple domains, can understand the context of a situation, and can learn and adapt to new problems without individual reprogramming. It has a much broader range of tasks and the same tasks can be accomplished in vastly different ways which require creativity and critical thinking.

Superintelligent AI, or Artificial General Intelligence (AGI), is an AI system that has surpassed humans’ level of general intelligence. It will require dramatically more intelligence than the average human, capable of thinking and learning independently within unknown domains. This form of AI has yet to be achieved, but is a constantly pursued goal among the AI research community.

In summary, artificial intelligence can be divided into three categories based on the complexity of the task and the type of intelligence the AI exhibits. Narrow AI is the most common type, and is used for specific tasks such as facial recognition and natural language processing. General AI is the challenge of creating an AI that exhibits human-level intelligence across multiple domains. Superintelligent AI is an AI that has surpassed humans’ intelligence, and is still being pursued.