Sunday, February 12, 2017

My ‘Artificial Intelligence’ dilemma

My daughter came back from the school the other day and asked me if my father and his father also used to work on computers. I think it was related to some activity they were working on in the school. I explained to her what my father and my grandfather did for a living, which by the way, was not computers.  As I went further in my explanation she seemed surprised and amazed that back in the ‘olden’ days, ‘people’ used to do things and not computers.

I could not get over my own thoughts even long after my daughter had stepped off once she had found her answers. I was intrigued by the thoughts of my childhood. There were not many devices or machines or cell phone apps to do things. Most chores had to be done manually. The most important point, in my opinion, was that back in the ‘olden days’ people had to ‘know’ how to ‘do’ things; unlike today, where people have to ‘know’ how to ‘use’ things. A finance person back then had to write up a journal, a ledger, a trial balance to get to the balance sheet. Today all that a finance person has to do is provide raw data to a software which throws out a balance sheet. Teachers are now ‘virtual’ and ‘online’; while signal lights, cameras and radars are the new age traffic cops. I wonder how many bankers now actually know how to write a balance sheet or how many students know how to write a sick leave letter or how many industrial workers know how to work on lathes and milling machines.

Slowly my thoughts strayed into the world of ‘Artificial intelligence’ (AI), the supposedly next big revolution. With the advancement in the areas of machine learning, self-healing, self - cloning-and-shut down; things are no more just programmatically done. They are now programmatically discovered, built, advanced, cloned, rebuilt and even destroyed. Humans are no more needed in many of the processes. Self-driving cars and trucks are replacing human drivers while ATM’s and teller-bots are replacing banking staff. Robots are replacing human workers in industries, in hospitals and even in armed forces. Maybe 20 years from now, people will have their own robots which will work on their behalf. Maybe ‘my’ robot might take my place as a programmer and program on my behalf; while a ‘farmers’ robot might do farming on behalf of its farmer owner and a ‘lawyers’ robot might argue in the court on behalf of its lawyer owner.

The more I thought about the possibilities of AI, the more endless that list seemed to me. I was amused, fascinated, intrigued, and also terrified at my own thoughts. The opportunities seemed limitless.

As I thought through, one question came up often and has remained unanswered in my head. Assuming that the intelligent humans will get into the task of building even powerful AI systems, what job will an average human do, say 20 years from now? Going by the rate of growth in the field of AI, most jobs that average humans do today, might be done by ‘AI-Bots’ in future. So what will a regular guy like me do?



On a lighter note -
I feel there is 1 job which cannot be take over even by the best AI-bots. Throughout the human history, there has almost always been 1 kind of people who have made their presence felt in almost every culture, in almost every civilization and in almost every country – ‘the politicians’. I think politics is one safe career that the next generation average humans should think about until the intelligent folks come up with something that the average folks can look into.



~Narendra V Joshi