First Type of Functional AI — Reactive AI

To understand the most basic type of AI that does not depend on any past nor has any memory stored, it is purely reactive.

Reactive AI is the most basic, simplest and first form of Artificial Intelligence. It is the oldest AI with very limited memory. Reactive AI acts in the present and is not based on any past experiences, meaning these machines cannot “learn”. It reacts with the predictive output from the received input. These machines are used for responding automatically to limited information. They do not have memory-based functionality so they cannot rely on memory to improve their operations based on the same. This type of artificial intelligence acts on what it sees. It doesn’t rely on any internal memory. They cannot function beyond the tasks assigned to them. These machines will react the same way each time if the input is not varied.

The decision process of Reactive AI is divided into 4 steps

Step1 — The first step is Commander — it analyzes the game report and looks for certain patterns in terms of situations. These situations may include enemies on the border, resource scarcity and lack of exploration.

Step2 — After detecting the situations, the Consultant generates general solutions or hints by consulting the genes together with an alarming level of the situation.

Step3 — In the next step, the Commander sorts the hints by the weights given to them in the gene and their alarm level.

Step4 — Hints are gradually converted into actual commands by the Micromanager

An example of the Reactive AI is the Deep Blue supercomputer designed by IBM in the 1980s which won a chess competition against world champion Garry Kasparov. It was programmed to identify the chessboard and its pieces while understanding the pieces’ functions.

It looks at the pieces on the chessboard, the way it is right now, and accordingly chooses from the next moves. It can predict, select and win with the exact moves. Its reactive mind indicates that it has no concept of past or future, it only acts on the present situation. They cannot create memories or use information learnt to influence future decisions — they are only able to react to presently existing situations. It does not learn or improve as it plays.

Netflix recommendation engines and spam filters are also examples of Reactive AI.