The moral Map of AI and the need for a value model for AI
The AI debate of how AI would be either very beneficial to mankind or very detrimental to mankind. And this reminded me of a discussion I had with another colleague of mine in the AI space, and his argument revolved around the fact that AI would exponentially learn of what to do in the human environment and start collecting facts about investing and would soon start to be a Investor better than Warren Buffet. My argument was that AI was a replay of the petrol engine model of the 1890s into 2020s environment.
What started as a discussion soon veered around to how would AI make moral sense of the world. There were far too many dimensions of human existence that the AI would not get starting from an infant who was totally dependent on his mother for existence to his middle age when his or her hormones and social standing dictated his/her behavior to old age when his/her experience dictated the persons behavior and one common thread we found throughout a persons existence on earth since he started to understand the world was the presence of a moral map which guided his actions, be it in childhood the persons old age and demise. Contrast that with an AI agent that will live from 2020s understanding data till around 2 centuries later in 2220s. We then discussed will the AI agent understand the world as Mother Teresa or as Adolf Hitler or similarly as Dada Idi Amin or Abraham Lincoln. There was silence in the discussion and the realization that the Human existence comes with a moral map of what is Good or what is Bad as a behavior in human society. In human society or human intelligence, you have extremes of human behaviour from Mother Teresa, who cared for people who were sick to Adolf Hitler who dreamt of a 100 year old reich. On another hand Dada Idi Amin who title when in power was "His Excellency, President for Life, Field Marshal Al Hadji Doctor Idi Amin Dada, VC, DSO, MC, Lord of All the Beasts of the Earth and Fishes of the Seas and Conqueror of the British Empire in Africa in General and Uganda in Particular'" to the other end of Abraham Lincoln who had failed in business before he became president and abolished slavery. These paths were varied for all these 4 people. We then turned to AI, would there be Abraham Lincoln AI or Idi Amin AI or Mother Teresa AI or Adolf Hitler AI. We look at the two extremes of Human Intelligence, from the depravations of Idi Amin and Adolf Hitler to the highs of Abraham Lincoln and Mother Teresa. Does AI have a moral map like humans?
What started as a discussion of the issues AI would have on future Humans, we came to the realization that AI did not have a moral map like Humans do and their existence is based on a button which starts to power the AI computer. And they are more similar to a petrol engine coming into the human society. The reason being the comparisons with a petrol engines development were similar to AI. When the petrol engine came into being, humans were used to horses for riding across the country and large distances, The horses served as an alternative to Human power, They could power vehicles that moved faster than horses and could be used to pump water and drive fans unlike humans who brought water from wells and rivers and fanned themselves when hot. Anything which required power could use a petrol engine. Just fill some petrol in its stomach and it used that to produce energy that will carry you faster than a horse carriage. This resulted in a changed VBE model which required roads and petrol stations for fueling the new cars with petrol. Gone were the Hay markets which were the feed stops for horses. You also got rid of horse manure and in its place was petrol exhaust from engines which dissipated in the air. We realized the issues with the CO2 emissions only later.
Similarly the AI are programs which understand data and its patterns and anomalies faster than human minds can understand. This gives the AI data sources for training with big data sets to understand patterns which is called machine learning. Once the AI learns a pattern it is able to spot anomalies in pattern and respond like a human would do , except that using the same AI you could have 1200 new AIs to handle human processes at the switch of the button. But what is left out of the AI intelligence is the Moral map of its environment. So yes the countries that build AI are building Large equivalents Computational power similar to the large petrol engines which gave power in the 1920 and yes there would be economic development which would follow with using Nvidia chips, but there was no moral map to help humans.
So the comparison of development AI will be similar to the way a petrol engine developed from 1877 till the cars and roads which followed. Where will these computation engines be used? I believe in areas where the human brain can be mimicked in the same way as human muscle was used with the developing power of the petrol engine. A human could not run faster than the horse or animals, but with the petrol engine they rapidly gained speeds which transported them faster than the animals they used for transportation. AI will be used in a similar manner to mimic human brain responses with the additional AI power.
This is a more accurate mapping of the way AI will develop.
The AI power will be used in areas where human computational observation is required to handle daily situations. What remains with humans is the moral dimension which AI cannot replicate. There will be no Mother Teresa AI or Abraham Lincoln AI as there will be no Adolf Hitler or Idi Amin AI. And you would not fall in love with AI intelligence. As with the petrol engine which required carriages and roads which then led to the development of road and carriage technology and also led to the demise of factories which produced horse carriages and hay farms which grew hay for horses to eat. There were Hay markets in many towns and even in Stockholm Sweden there is a place called Hotorget which was the hay market in earlier horse transportation times. The business which declined were the breeding of horses for transportation and the hay markets which fed horses, and this gave rise to petrol extraction and car building industries. The transportation now shifted from horses to cars. Similarly with AI the jobs which will be affected will be placed where AI can be deployed to work, and we will see driver jobs, ticketing agent jobs and such jobs which require a small amount of human thinking moved over to AI agents. The moral jobs will still remain with humans, Imagine an AI barber who charges 2 USD for a haircut and will accept all suggestions you have of keeping the front of your hair long and trimming the hair at the top of your head or whatever style you have in mind. What will be the value model for AI as has happened for petrol engines? Ever new technology needs a VBE model (Short for Value Model, Business Model and Ecosystem) to survive and that will be true here too. The VBE model is explained here and has been operating for the last 150 to 200 years. AI will follow the same pattern.
But the plus point with the AI development is that moral map will still remain human.