When it comes to judge something, our processing speed reaches its ideal peak which as per the facts is around 1017 operations/sec. But, is it necessary to judge everything and are we really accurate in classifying objects and events in good and bad category? And how this classification works? These are some questions which are necessary to be answered prior to develop an AI agent which is going to be an extension of human thoughts processing. Here in this article I am explaining these the same.
Starting with the basics, thinking is one of the primaries for living beings. Since it is executed in human brain and our brain follows cognitive psychology, which views the brain as an information-processing device. In the article, AI: A Theoretical Intro, I mentioned the three basic ways to understand thinking as an operation of our mind, which are – through introspection, psychological experiments and brain imaging. Psychological Experimentation also provides Learning from the past as a method to think. But before that we need something in our knowledge database which we can call the past. It has been found that we as a human store objects, relations and events in a labeled-classified manner in our mind which we can trace back using introspection or simply remembering and retrieving data from our brain. It means that if we see a person, its image get into our brain and brain returns all the labels and classification, based on our past experiences in accordance to the situation or circumstance. This way we gets a clear picture of the person/object/event. Now, this is a brief about what brain responds after visualization or sensing something. But, where is Good or Bad categorization? Is it a stored response by brain or instantaneously processed? Well the answer is both. Primarily it is calculated instantaneously which then stored for future reference.
Processing data is one thing while categorizing it in good and bad, needs an intro of these terms. As these are relative and complementary to each other, you can not feel goodness if there is no badness and vice-a-versa. We all have our own references of these feeling elements. Those references are also known as standards. Like, when we are young, our parents used to say that speaking lies is a bad thing, and some of us followed it and taken it as a reference. While you might also have a feeling that if a lie is going to help people then go for it. It is good! Again, this can be one of our standard. These type of standards helps us to judge ourselves when we speak. And we usually apply these to others for judgement. This is where good and bad terms arises for us where past and community standards are decidability factor for good and bad for us. Since, we know that these types of standards might not be universal as human thinking is more personal and situational rather than universal, thus can we judge efficiently and in a right way?
Again, doing this efficiently and in a right way are two different things. For an AI agent, both are required, as no one wants to have a system which doesn’t do right things or is in-efficient. So, let me extend this by using the term rationality. It is all about a system which does “right thing”, given what it knows. And in some books these right things are considered as universal in nature. For example
Let me put a circumstance in-front of you. Suppose you are at a place in the middle of nowhere and have a gun with you(for self defense). Now there are also two people say A abd B where B is running after A, pointing its gun. Situation is such that, B is going to kill A, if you don’t intervene. There will be no talk, and you have to take a decision instantly! Either you can use your gun to kill B in order to save A or no interference and let A die. What will you do in the following cases?
Case 1: A and B, both are Men.
Case 2: A is a Men and B is a women.
Case 3: B is a Men and A is a women.
Case 4: B is a legal officer and A is a terrorist.
Case 5: A is a legal officer and B is a terrorist.
Answers will vary person to person. But one thing is very much clear that, more information will lead to more efficiency to judge the situation. And degree of goodness and badness will depend upon your standards. But rationality here is to neglect everything since, that gun is defined for your self defense(first of all knowledge) and not to help other as trying to help may rise uncertain consequences. I know, some of you think it as weird but I am not saying that this will be a right choice for you also as we are humans and have a free will to think and execute, but this article has intentionally been written to understand how an AI agent should work in these situations of dillemma. These can only be overcome by rational thinking rather than humanly thinking.
Leave a Reply