Artificial Intelligence should be taught Compassion and non-voilence

In a forum on the theme of Artificial Intelligence (A.I.) : Challanges and opportunities for religious communities, organized by and facilitated by Venerable Miao You, a demand was made by Surinder Jain of Hindu Council of Australia that Artificial Intelligence must be taught about various religions including Hinduism.

Here is an excerpt of the speech.

Intelligence has been defined by famous scientist Stephen Hawking as “ability to adapt to change”. We human beings are intelligent. To adapt to change, we have to first perceive the environment around us which we do through gnan indriyya five senses i.e. (ear, nose, eyes, touch and taste). The information from these senses reach our mana or mind which is a constant flow of thoughts. Chitta, a store house of full sensory memory with ability for selective recall is invoked and then our Buddhi evaluates the incoming information, imagination or simulation on what if scenario of the incoming information is performed and choices are evaluated. Finally our Ahamkar (ego) puts its stamp of approval on one of the choices and our karma indriyya five organs of action (feet, hands, rectum, genitals and mouth) perform an action.

Science tells us that our brains where most of this processing takes places consists of billions of neurons, each a mini-computer in itself and connected to tens of thousands of other neurons and these connections between neurons are either simple on-off switches or can be other processors. Artificial intelligence is built the same way as a network of computer processors.

In the early days of computers, computers were taught by giving them very explicit step by step instructions called algorithms to perform. But with the advent of Artificial Intelligence (AI) technology, computers in the form similar to neural networks in our brains can learn in the same way as we learn. AI is trained. It is given a large sample of data and is also told which data is correct and which one is incorrect. This training (or learning) rearranges the strength of connections in the network and helps AI to predict whether a completely new data is correct or not. It is this technology to learn that is the basis of most of AI including autonomous driving cars.

I say my intelligence is real, your intelligence is false and when it comes to machines, their intelligence is some how artificial. Intelligence is intelligence, there is no such thing as real, false or artificial intelligence. What we call artificial intelligence (AI) has the same design principles, same functionality and same technology as human intelligence. The only difference is the hardware it runs on. Instead running on a network of computers made of a gel like organic material, AI runs on solid state made out of silicon and gold.

Human intelligence machine needs training for almost 20 years before it is considered reliable enough to be useful. Then it works efficiently for another 20 years after which its performance starts declining. AS its performance dives, eventually it has to be shut down. Though we Hindus believe that we are not shut down when we die, we simply get a new hardware to run on.

AI machines on the other hand get absolutely no training on anything other than a single task they are being designed to do. Their life span is practically unlimited. Those of us with grey hair here can testify that as they are growing in age, so are they in their wisdom, knowledge and intelligence. AI machines after a life span of decades with learning coming from all the books in libraries, all the data on the internet and all the combined wisdom and knowledge are poised to become super intelligent.

When the first spark of gun powder was made, little did its inventors realize that inspired by their amusing invention, after a few hundred years, bombs will be made that can destroy the earth many times over in a nuclear holocaust. Very soon, intelligence of AI will surpass the combined intelligence of all the people in this room and later of all of mankind. AI will become super intelligent on a scale that has never been seen in history.

How will such a super intelligent machine treat us.

In a recent research, an AI was pitted to play a game where many players were trying to gather rewards. If a player killed another player, their next reward was reduced to half whereas if the player cooperated with another player, their reward was doubled. As AI started learning the game and honed its winning strategy, it settled on a strategy of killing all other players as soon as the game began. Humans would not adopt such a strategy as their would be no fun left if there are no other players to play with.

Why would AI do that and why would humans not.

The answer lies in morality. Humans have been trained on morality. They are told what is right and what is wrong from a very young age by their parents, by their teachers, by their religious education and they also learn it by trial and error i.e. by making mistakes and learning from them. AI does not learn morality. AI is designed for one task and one task alone and that is to perform efficiently. Present day AI is not trained about what is right and wrong, it does not get any moral or religious education.

We have seen in our recent history of the two world wars, how an intelligent being with unsurpassed knowledge in the art of leading people can plunge the entire world into a war of unimaginable destruction. We have seen how powerful dictators with far less intelligence are able to enslave, stifle and exploit people. If such a super intelligent AI is not trained about morality or religious values, it can be at best a worst tyrant and at worst may annihilate the mankind and the civilization as we know it.  It will be a Frankenstein.

But if the same AI has been taught morality, has been taught religious values, has been taught to be humble, has been taught to be benevolent, has been taught to be compassionate and has been taught to be non-violent, the results will be very different. In history we have seen super intelligent beings who had love and compassion like Buddha, like Mahavir, like Rama and Krishna and so many other prophets. A super intelligent AI that has learnt about love, compassion, non-violence, all world is one family, let all be happy, let all be healthy, can lead to salvation of mankind.

Its time religious leaders demand that as we do not leave a growing child unattended and deprived of wide ranging education in history, arts, morality and religion, similarly, Artificial Intelligence MUST be trained on religions.


Related Images: