AI and the future of Jobs

Sejuti Banerjea

“It is extremely easy to find people who speak pleasantly. But it is rare to find people who speak and hear true words even when they are not pleasing to hear.” From the Ramayana, as quoted in Rajiv Malhotra’s book Artificial Intelligence and the Future of Power (referred to as AIandPower through the rest of this article).

It seems necessary to start with the above words for the following reasons-

First, this question is relevant to every single person in the world; everyone that is who works for a living and everyone with a loved one who must work for a living.

Second, it is relevant to every single person that wants peace both in the immediate surroundings and also through the rest of the world.

Third, it is important to answer this question responsibly, so we can prepare for the worst while we continue to hope for the best.

The truth is, we have not done a good job of living responsibly. That’s why we have a swelling population (“the world population is forecast to increase to 9.7 billion in 2050 and to 11.2 billion by 2100” Pg 98, AIandPower) and significant concentration of wealth in a few hands. 

Malhotra’s book gives us some astounding numbers: “In the Global Wealth Report 2019, Credit Suisse Wealth Institute indicates that the top 1% of the world’s richest people own 45% of the world’s wealth…According to one report, the world’s 2,000 billionaires have more wealth than the bottom 4.6 billion people combined, and the richest 1% have more than the combined wealth of 6.9 billion people.” Pg 88-89, AIandPower

These two problems are at the root of the issue at hand and threaten to exacerbate its negative effects. But because the extent of the problem isn’t obvious, it is necessary to elaborate.

Just imagine the consequences of increasing mechanization/automation/robotics in a world where the number of people looking for jobs continues to increase while the ones owning wealth and resources remain few in number. Obviously, the job searchers won’t be the ones with the resources to buy these machines. They can be afforded by the lucky few that haven’t been displaced by machines (until they too are displaced) and by wealthy industrialists who use them instead of people. Because as machines become more capable, their cost of ownership falls while wages naturally increase over time. So it makes financial sense to use more machines. 

Artificial intelligence (AI), the phenomenon of “smart” machines, is mankind’s latest innovation. Human beings by nature are innovative and inquisitive. We are always trying to test the limits of our capabilities, both physical and mental. But over the years, we have already mastered many physical activities. Through innovation and interaction with the material world, we have devised systems and processes that can make our lives much easier. We don’t, for example, go to a river or lake to collect water. We have pipelines that deliver it straight home. Nor do we walk miles, the way our ancestors did. We have appropriate vehicles to do the job. So, in this way, we have delegated many routine activities to machines in order to make our lives easier.  

Just a couple of generations back, receptionists and telephone operators were a must in every office. But while some offices still hire receptionists, telephone operators have become obsolete. Many line managers work without assistants, because so many functions are now taken care of by machines. And even more recently, we are all working from home and ordering food, groceries and other necessities with the click of a button. So automation reduced the burden of physical work necessary for survival and made our lives easier. 

But in the past, “machines were not replacing judgment, intuition and creativity” (Pg 98, AIandPower). AI in essence, is doing all this and more: its goal is to also take over a lot of mental work. In fact low-end white collar jobs are what machines can quickly take over. Actual robots replacing physical laborers require more work to build, customize and deploy, so they will only be deployed over time. For example, Amazon uses robots only in its warehouses and for some deliveries, but it uses AI across many other operations.  

So in a sense AI is not doing anything new. It is part of the changes that have been happening over the years. But because the changes of the past happened so slowly, without disrupting our lives, we didn’t pay attention to them. AI is going to change all that because technology adoption is very fast today and its advantages show up immediately in corporate profit and loss statements. 

On a personal level, we all know how fast mobile phones have caught up with us. These devices weren’t required by anyone a few years back. But today, we’ve discovered how convenient it is to have a digital connected device in the palm of the hand. So they are our constant companions. 

The phones and other computing devices connect us to the worldwide web, which is dominated by American technology companies like Google, Facebook, Twitter, Netflix, Microsoft, Apple and several others. So the way we interact with the technologies provided by these companies creates a digital blueprint of our lives. It contains valuable personal information about us and the way we do things. 

“The large digital platforms are the key engines today that mine the data and curate, analyze and apply it. Social media brings together disparate parties to interact and captures their data in the process.” Pg 70, AIandPower

We may not consider this information valuable because we are personally just small entities in relation to the world in which we live. But when millions of us hand over this data, it can be fed into advanced machines with very large-scale processing configuration called neural networks that mimics the way our own brains work. So these machines can identify our behavioral patterns. 

And along with a basket of other technologies like natural language processing (NLP), nanotechnology, facial recognition, quantum computing, etc, these machines are getting more and more intelligent. So today, they can produce articles or correspondence, compose the most soulful music, provide companionship and advice to the aged and recognize you from a crowd of over 500 people, even if you’re disguised in some way, and even in the dark. They can use medical data super efficiently and speed up drug discovery by years.

Also, while there’s a limit to our personal faculties, it’s much easier to add compute power to machines, so they can do even more work. Since they also never forget anything, they are really hard to beat. In short, machines can do most things better than humans, simply because they are so good at reverse engineering the conditions and processes that lead to certain outcomes and then replicating those conditions and processes. With machine learning, machines can train themselves to get better and better at what they do, as they process more and more data.

So there’s no doubt that without intervention by governments and broader awareness of the issues by people at large, machines will take over most jobs. But there’s less clarity on how this whole thing will unfold. Malhotra’s book quotes several sources and highlights some possible scenarios-

  1. A quote from a Bain & Company report reads as follows: “In addition to job loss and wage suppression, automation may also increase income inequality by increasing the share of income going to profits vs. wages.” The report further says that “Capital ownership is already highly concentrated…Because capital ownership is tilted toward those already in higher-income brackets and also much more narrowly concentrated than income, this shift toward capital income is likely to contribute to rising income inequality.” Pg 89-90, AIandPower
  1. “AI’s new jobs will not be located where old jobs are eliminated, but wherever the AI industry’s innovations and implementations are located…The consumers who benefit from cheaper goods due to automation will be scattered around the world, whereas the communities that lose manufacturing jobs will be locally concentrated.” Pg 91, AIandPower
  1. World Economic Forum’s (WEF) 2020 report on global risks notes: “automation is forecasted to hit low-skilled workers and women the hardest. Societal divides could also widen between rural and urban areas in developing economies, and between smart and non-smart cities in developed countries.” Pg 91, AIandPower
  1. The low-wage advantage of workers in some areas is currently being eroded because digital platforms reduce workers’ bargaining power.” Pg 92, AIandPower
  1. Economists, politicians and commentators keep reassuring us that education and retraining can solve the problem” But according to Daniel Susskind: “the traditional response of “more education” is likely to be less and less effective as time rolls on” Pg 100, AIandPower
  1. While automation will enhance the capacity to produce sophisticated products, the lack of sufficient consumers that can afford to buy those products will lead to an underutilization of production capacity. Extreme economic disparity eventually causes a downward spiral in demand and could precipitate a deflationary economy that contracts.” Pg 94, AIandPower
  1. Such a profound disruption in the labor market will trigger social unrest… divisions will tear at the social fabric and undermine human dignity. Widespread unemployment, combined with increasing inequality, will lead to social disorder, political unrest, and even threats to the sovereignty of many nations. Pg 95 AIandPower

India being a country with a huge and growing population and tremendous concentration of wealth, will be one of the hardest hit. This is more of a pity because Indian tech workers are helping build a lot of the foreign intellectual property driving these changes.

Sejuti Banerjea
Sejuti Banerjea

Financial analyst with interest in religion, philosophy and spirituality

Leave a Reply

%d bloggers like this: