Table 3.

Functionality- and technology-based classification of AI.

AI Type	Properties	Ref.
Functionality based
Reactive Machines	The earliest AI type, with the most limited technological capabilities. Operate without the need to have a memory base. Mimic the human brain’s capabilities to react to many sorts of scenarios. Cannot store experiences and use the results to predict future responses.	[61]
Limited Memory	It merges the ability to react and learn from previous information. Has many learning algorithms; the collection of datasets helps AI systems with analyzing current situations and learning from past experiences to make decisions in the future. The best example of AI-limited memory systems is the fingerprint scanning machine.	[62]
Theory of Mind	It is designed to perform a broad range of analyses—of thought patterns, emotions, and belief systems. It will help robots better comprehend people and understand the different elements that impact their thinking.	[63]
Self-Aware	Theoretical systems remain under development. Sophisticated systems comparable to the human mind; will have self-consciousness and be self-sufficient. It is still uncertain how long these AI systems will take to evolve.	[64]
Technology-based
Artificial Narrow Intelligence (ANI)	Now, ANI is the most widely used category of AI. ANI is capable of doing one or two jobs at the same time. It takes advantage of the training data, and experiences gained from past situations. Currently, all models of AI may be classified as a part of ANI, from fundamental methods to advanced algorithms used by computers to make decisions. ANI is considered a weak AI system because it operates within a limited set of parameters	[65]
Artificial General Intelligence	This is linked to the Theory of Mind for functionality scientists are still working on. The goal is to design devices that can independently create connections throughout several domains.	[66]
Artificial Super Intelligence (ASI)	It is designed to execute activities and make choices better than humans, who demand massive quantities of memory. Machines with a large amount of memory, a higher processing speed, and a faster rate of intelligent decision-making would be able to perform complicated tasks with ease and in less time. They will be able to make complex decisions impacted by various circumstances in ways they have never been able to before. It is still under development.	[67].