Classifying AI on Functional Basis

AI is basically classified into 4 types based on their functionalities. It has the capability to learn new things by itself, demonstrate, explain, and can advise its user.

In AI, the word Artificial is defined as ‘man-made’ and intelligence is defined as ‘thinking power’, hence AI is “a man-made thinking power”. AI is a combination of Reasoning, learning, problem-solving perception, language understanding, etc. It is replicating human intelligence by solving knowledge-intensive tasks.

We will start with the basics in this article by explaining AI and its types. Eventually, as I will be posting daily, we will see each topic in detail. You may link them one after another for a thorough understanding of all these AIs.

AI can be classified into 4 types as per functionalities:-

We are currently thorough with the first type and in progress with the second one for perfection. Eventually, the next 2 are in a theoretical state yet. We may call them the next stage of AI.

Reactive Machines

This is the first and simplest level of AI that performs basic operations. Reactive AI is programmed in such a way that based on the input received, it predicts output. Reactive machines never store inputs nor learning.

For eg — it takes a human face as an input and outputs a box around the face and identifies it as a face.

As it was the first type of AI in the history of AI development, there are certain limitations associated with it. They don’t have the ability to perform tasks other than only the ones which they are designed for.

Limited Memory

Limited memory AI is widely used in today’s world. In the limited memory type of AI, it stores the previous data which can be used for making better predictions. It can perform complex classification tasks with the use of past data like historical, observational data with pre-programmed information. It is capable of building experimental knowledge by observing actions or data. Machine learning models like Reinforcement learning, Long Short Term Memory (LSTMs), Evolutionary Generative Adversarial Networks (E-GAN)help to achieve this Limited Memory type.

For example, limited memory AI is used in autonomous vehicles to observe other cars’ speed and direction, helping them “read the road” and adjust as needed.

Theory of Mind

Theory of Mind is the type of artificial intelligence that is at the beginning phase and can be seen in things like self-driving cars. This AI interacts with the thoughts and emotions of humans. Machines will be able to understand and remember emotions and adjust behaviour according to these emotions as they interact with people. Machines in this type of AI will be capable to make decisions that are similar to humans. Yet achieving the theory of mind AI is in process.

Self-aware AI

This is the most advanced type of artificial intelligence. Machines will be self-aware of their internal emotions and mental state. When machines can be aware of their own emotions, as well as the emotions of others as humans do, they will have a level of consciousness and intelligence. This type of AI will have desires, needs, and emotions as well.