What is Artificial Intelligence(AI)?

Artificial Intelligence-The Exact Definition

By ANONYO SANYAL

Artificial Intelligence(AI) has been hot for around past a few years now, but the concept was familiarized by Alan Turing in the year 1950. In one of his papers called "Computing Machinery and Intelligence," he first introduced the concept of  AI indirectly. He mentioned a test famously known as "The Turing Test" where he discussed how to assess the intelligence of a machine. Before we go into the details of Turing test let us first talk about AI as a whole.

AI- The Definition

AI can be defined in various ways. Basically, if we say in a nutshell, AI is simply 'mimicking' the human intelligence by some machine. But this mimicking can be done in four broad ways v.i.z. 

1. Machines who think like humans.
2. Machines who think rationally.
3. Machines who act like humans.
4. Machines who act rationally. 

These are the four ways how Peter Norvig and Stuart J. Russell has classified AI in their celebrated book called AI-A Modern Approach. Does AI only means mimicking humans? Can't we mimic animal intelligence? The question is what exactly is Intelligence? A dog can say it is more intelligent than humans in some aspects. Who will judge that? A cat can also say the same as well. So why aren't we mimicking that intelligence? Frankly speaking, we do mimic them. There have been some robots which mimic wild animals. Say a tiger. A robot which looks exactly like a tiger is built and is given the intelligence of a tiger and is put in a pack of tigers. Cameras are there in the eyes of those robots which help us understand how Tigers live. Those are also intelligent machines. They are not Machines who act like humans but they are Machines who act like Tigers. Is it not AI?

Secondly, people get mislead by the word "rational". People generally think rational thinking or acting rationally means not to think or act crazy. NO! Here rational has a whole different meaning. For example, you say sentences. Even though you know all the grammatical rules but you might fumble and even speak sentences which are grammatically incorrect. We say such behavior is irrational in the world of AI.  

In fact, it is quite interesting that the definition of AI changes with day to day life. To understand this let us dig the soil for a while. David Neil Laurence Levy is a British International Master of chess had a bet saying that in the next 10 years(with respect to 1968 when he placed the bet) no chess playing machines could defeat him. Luckily he won the bet, but back then AI meant this. But now we fairly overlook that a chess-playing computer game is also an example of AI. In fact, there are many chess playing machines whom one can never defeat. They are intelligent as well. Now the later is also an example of machines which act and think rationally. 

z

The Idea Came in 1950 Then why all the work is being done now?

The reason is pretty simple. Back then in 1950 mankind had not achieved such computational power as of now. Memory is cheap, Processors have become faster and above all the increase of computers in every sector has resulted in enormous computational workload for humans. So if humans could share those loads with machines who are smarter than humans would definitely make our lives simpler.

What about Turing Test?

Turing Test. Ah! The real stuff. Let me explain it to you all with an example. Suppose you are a judge and you are given a task to identify humans and computers by only a Q and A session. Pretty simple? No there is an added constraint. You cannot perceive them with any of your senses. They are hidden from you. Say for example there are two samples-A and B. You ask questions to A and the replies you have got has convinced you that A is a human. Then you do the same with B and again you are convinced that B is also a human. Now you have the report card in front of you. You see that A is a human. You were right. But then a very disturbing thing happens. You see that B was actually a computer but you were convinced that it was a human. You were wrong. This test is known as the Turing test.

The outcome is that you say that machine B is intelligent or more specifically Artificially Intelligent.

Finally...

Finally, the definition of AI is relative to what system you make but the definition is classified in four main sub-parts as said above. 

Think Philosophically

If some day humans can make a fully Artificial Intelligent system which not only acts like humans but are rational as well can't we say we will have the same power that God has: as the Almighty made us.

Think about it and comment below your thoughts.