Of late, AI has become more ‘popular’ in the public discourse. I often get invited to deliver presentations and public lectures on AI from an interdisciplinary perspective. Therefore, I also get to hear from a lot many experts hailing from different disciplines. There is no consensus on any aspect of AI. However, one thing comes out very strong from those who try to articulate a positive impact of AI on jobs. My perception is totally different, for I have reasons to believe that there will be no job, whatsoever, in the contemporary economic sense, in the AI future. But, who will need jobs, as they exist today, in the AI future? Probably, none!
There are some, if not many, scholars who believe that numerous jobs will be created if many are going to be lost because of AI. In this context, Bullshit Jobs: A Theory, by Prof. David Graeber, becomes relevant. In the transient period i.e. from artificial narrow intelligence (ANI) to artificial general intelligence (AGI), and possibly, then, to superintelligence, some jobs might be created in order to compensate for the jobs that are going to be lost because of AI; it will be done primarily as a desperate attempt to sustain the current economic system which is ultimately going to be overhauled by AI for the reasons that I will ponder over in another post. However, the new jobs thus created will not be meaningful or relevant. In fact, there are many jobs, today, that the society does not need; but they still exist. Existence of something does not always correspond to its worthy need in the society.
The reason that most scholars advance in order to alleviate concerns of AI led unemployment is expressed as reference to the consequences of earlier three industrial revolution. AI, in the economic context, is often referred to as the fourth industrial revolution. According to these scholars, automation caused by the first three industrial revolutions took away some jobs but it did create others jobs, as well. Without going into the normative assessment and merits of assertions regarding the nature of the jobs that were created by the earlier three industrial revolutions, it is pertinent to mention that this argument is inherently flawed because the fourth industrial revolution is not even remotely similar to the earlier three. Why?
From steam engine to electricity to computer, the man could do the things that were beyond his physical capabilities. The man could mechanise and automate production. The said automation was completely subservient to human control, in principle and practice. In other words, every machine or system until the beginning of the fourth industrial revolution followed detailed instructions provided by humans to achieve the desired result.
But, what AI seeks to do is to take decision-making and its implementation away from the domain of humans to artificial intelligent systems. It implies not only complete automation of every aspect of human existence in the AI future but also humans, for every practical reason or otherwise, to the exclusion of their own ‘meaningful’ intellect, relying completely on decisions made by AI. In fact, the said ‘reliance’ will only be notional, for implementation of AI decisions, with humans as subject, will also be in the domain of AI.
The aforementioned situation becomes more intense when one considers the work done by Elon Musk’s company, Neuralink. Neuralink is in the process of developing machine-brain interface. If human intellect can be augmented and finally replaced by more computational power of systems that are capable of being intelligent in any manner whatsoever, human beings will not be what they are perceived to be.
In closing, those who advance the argument that if the earlier industrial revolutions did not essentially take away jobs, the fourth industrial revolution will not either, should reconsider the substantial implications of AI that is advancing at a very rapid pace. The very underlying idea of AI is to have a system that is capable of having its own independent and non-human intelligence. Anything to the contrary would be counterintuitive and defeat the very purpose of AI. Jobs would anyway become redundant if AI brings along the expected exponential economic growth, with high productivity and increased output. There will be enough for everybody.