The socio-economic impact of Artificial Intelligence

Amid all the excitement about artificial intelligence, one thing that has being occupying me the most (as a person who could be replaced by a machine) is the impact of automation on the broader economy. These past couple of months, this topic has been getting a lot of media coverage. I recently read a Bloomberg article regarding a machine which could perform thousands of man-hours worth of legal work in minutes. Think of all the paralegal jobs at stake. More recently, in India, reports that the IT outsourcing companies and banks (which are some of the biggest recruiters of college grads in the country) are eliminating positions which can be automated, ought to have given everybody from the government to the parents of soon-to-be college grads nightmares.

It is certain that we are at the cusp of an AI revolution. Computing power is growing exponentially every year, and we have also been able to cross the biggest barrier to AI so far- bounded rationality of humans. AI was thus far limited because humans could not possibly predict and build into their programs every eventuality. Something always used to be left out in extremely complicated programs. However, with enough computing power, we now have the resources which enable computers to learn from vast quantities of data which would make no sense to humans. Thanks to “Big Data”, machines can learn from their environment without each input having to be programmed by a human. Essentially, a machine could operate as a human brain- learning through experiences just as a baby does. This was on display when Microsoft’s bot was exposed to the usual vitriol on Twitter and it too began to behave just like one of the extremist trolls. Essentially, this is what a human child would do- learn from the environment it has grown up in and then become a reflection of it. One would argue that it is still impossible for a computer to replicate all the functions of human decision-making, such as relying on gut instinct. It isn’t impossible that using Big Data, a machine could figure out patterns which aid in decision-making- what we attribute to our so-called “gut instinct” (since we can’t explain it).

In any case, there is no doubt that computers will be able to perform most tasks which people perform in their jobs. Most humans rely on a defined set of heuristics while working. This is more prevalent than what many people think, which is why some people who believe that they are safe from the rise of AI are set for a rude awakening. Machines will be able to do everything faster, with fewer mistakes and cheaper. Technology only gets cheaper over time, whereas human labor does not. The future where machines take over most tasks, leaving billions of people unemployed, is not far away. I will attempt to pen down my ideas of what that future might look like from a socio-economic perspective.

First of all, let us consider which professions are more vulnerable to disruption. Any job that has a well-structured way of going about it (with few variations) and doesn’t require much subjective judgment is in a precarious position. Some jobs are like that, but they rely on some forms of experience as well. Those can be automated too. Robots are already capable of doing things right from serving food to surgery. However, there are jobs which do not cater to a strictly rational or functional human need. They are primarily in existence to fulfill our emotional needs. These range from anywhere between a psychologist to a professional football player. I can’t fathom that in the near future, such jobs can be carried out by robots. The same can be said about most artistic pursuits. I am unable to perceive the day when machines paint, dance, act, write elegant poetry, make movies, or create new recipes. Would it be possible for the super-intelligent machines to also acquire a sense of humans’ tastes and preferences- and if so, influence them just the way companies do today through smart marketing campaigns and propaganda? I hope not. I am not sure, either, of machines being able to innovate and invent stuff which cater to the ever-growing list of human needs which are not rational (such as the need for attention- which is taken care of by social media). For now, the people engaging in such professions seem to be safe. But people forced out of other jobs will naturally gravitate towards such professions- and that will cause their earnings to drop sharply. That said, at present we still spend much of our incomes on base needs such as food and housing (which could be catered to by machines)- and this becomes more prevalent the lower you go along the income spectrum.

It is said that the true cost of something depends on the human effort invested to create or acquire it. That is part true. I believe that when the time comes when most of the goods and services we use will be produced by machines, prices will fall dramatically for most items. Consider the case of self-driving cars and Uber’s attempt to master the technology. At present, each ride constitutes the following costs: fuel, the cost of the car (including maintenance), profit, and the cost of the driver. The rise of the self-driving car will eliminate the labour cost- which is the biggest component of the overall cost of a ride- completely. Additionally, the car will be cheaper and easier to maintain. It is true that labour may not be such a big component of cost in other industries, but it is a major one. The car itself will be built by robots (which will be made by robots in factories built by robots). At present, every component in every transaction involves compensation to somebody (a person). If all the tasks required to deliver goods or services could be done by robots, then there would be nobody to compensate.

Let’s delve deeper into the factors of production (land, labour, capital, technology, entrepreneurship) as taught in macroeconomics. We have already talked about labour. Let’s move on to entrepreneurship. In a capitalist system, the entrepreneur usually gets one of the highest payoffs from any endeavor (after investors in most cases, and definitely higher than labour). The entrepreneur takes the majority of the risk involved in any enterprise. To devote or not to devote his time and efforts to any venture, and the day-to-day decisions associated with the venture, are all dependent on his judgment. For example, whether to continue a steel manufacturing enterprise or not would largely be decided based on the entrepreneur’s evaluation of the supply-demand situation and the long-term prospects for the same. If computers are able to make far better decisions regarding this based on their vast troves of real-time data, then they should be allowed to. As we know, capitalism can lead to great economic waste due to under- and over- production and irrationality on the part of businessmen. If in the future, computers are able to communicate and share information with each other to make better decisions, the entrepreneur could prove to be extraneous. In that case, the compensation for that factor of production should come down too. Of course, it would be impossible to delegate all the aspects of decision-making to machines, but to a last extent, this would happen.

Capital. This is a tricky one. Capital will always be measured in terms of money and humans will always own all the money in the world. But as we know, ownership doesn’t necessarily mean control. People give their money to others to do something productive out of it and make their money grow. Even as we speak, such capital allocation decisions are increasingly being made by robots. Robo-investing, algo-trading — these terms would ring a bell. Again, not all capital allocation decisions would be left to robots, but many would. If that happens, management fees ( a major component of the compensation that goes towards capital) would reduce drastically. If managed properly (a big if), a financial system where decisions on capital allocation are made by robots would have a lot less volatility than what is currently present. With reduced volatility comes a reduced need to compensate capital providers with greater return. Thus, another factor of production might get a lot cheaper.

This leaves us with land and technology. Land, in my opinion, ought to include all natural resources used in the production process. Obviously these resources are finite the way things are. It is very possible that what we can provide for may actually be limited only by the amount of resources available on the Earth. Of course, the harnessing of these resources depends on the technology which we have available. Which leaves technology and innovation as the one factor of production which I feel would still largely be left in the hands of humans. I just don’t see how robots could go about the innovation and invention process especially given that so many fortuitous inventions have been results of mistakes or irrationality on the part of the inventors. I may be wrong about this, though- since I am a layman when it comes to the technicality of AI and computers. However, it must be noted that the price of technology is also largely a function of the compensation provided to the humans who created it. Once a technology is sufficiently developed and democratized, its cost only equals the marginal cost of deploying it. If even some of my earlier opinions hold true, then most people who currently aren’t in technical fields will be unemployed sooner or later. If research and invention is among the last fields where humans are needed, people will flock to it. Once that happens, the price of the labour will fall due to excessive supply and competition. So, technology too will become a lot cheaper.

Regarding the finite nature of resources we have, we are already trying to push the boundaries of that constraint. People like Elon Musk and Jeff Bezos are already making plans of mining in space and colonizing other planets. Since I have no capacity or the audacity to make predictions of how these endeavors will turn out, I am leaving resources as a question mark.

Since it seems that four of the five factors of production will become dramatically cheaper, it would make sense that the prices of goods and services in the future would only move in the same direction.

Now that it is established that there will be mass unemployment, yet at the same time, things will become far cheaper, let us see how this might impact society and the way it operates today. Does this scenario mean that even a penniless person can now afford anything he wants? Well, since I doubt that everything will actually be free, it’s unlikely that this would happen. However, his misery would be reduced by a great deal. The benefactors on whose goodwill he relies will now be able to buy a lot more with their money and I’m hoping that they would be more generous with it too. All these changes do raise the question as to how the concept of wealth, consumption etc. will evolve. I think we are already seeing this in action even today (though in a less prominent way than what might be the case in the future). There are already millions of people in the fortunate position of having all their material needs satisfied. What do you do if you are insanely rich, bored with just consuming more of the same, and need new ways to subtly flaunt your wealth? You spend your money on unique, tailored experiences. You try and do something which nobody else will be able to replicate and you’ll be able to tell that story for days, until somebody else matches you and you are off in search of your next pursuit. The same thing may continue. Sergey Brin is spending some of his fortune on a space ship. Others buy islands and so on. Even though philanthropy has caught on, the rich still barely engage in it (many would argue).

How, then, would we able to provide for those without wealth and who have been robbed of their income as well? Money has to flow from someone’s pocket into another, and if the rich don’t contribute voluntarily, then they will be forced to do so. As more and more changes come about and an increasing number of people are affected, there will be a lot of flux in society. People rarely respond well to change. I hope that it does not come to the point where there are mass riots and calls for the heads of the rich. Some transfer of wealth will have to take place for there to be continued peace, be it through taxes or donations or whatever. It is clear that society will undergo a huge upheaval over the coming decades (even Jack Ma has said as much). How this will be managed depends a lot on politicians and the rich. So far, they have managed to keep simmering tensions regarding rising inequality subdued.

Will the dynamics of capitalism change completely? If everything is nearly free, suddenly entrepreneurship will not involve much of risk or opportunity cost. In that case, the current system of the owner taking all the residual profit after paying all claims, will not last. If, as we discussed, machines make most of the decisions of production and capital allocation, would that lead to a return of the planned economy where all economic decisions will be reduced to complicated assignment problems wherein the objective is to satisfy as many people as possible within the given resource constraints? We’ll just have to wait and watch. In that case, one complication which might arise is with regards of decision making for machines- whose needs and desires get satisfied first and based on what criteria? At present, we know the answer- the person who can offer the most money. Will this change and equitable distribution be carried out, or will the existing order prevail?

I’m aware that this write-up raises more questions than answers. If I stumble upon the latter, I’ll be sure to let you know :)