One of the beliefs journalists and some who should know better is that AI will replace all labor. I find that idea so absurd that I don’t understand how anyone can believe it. I want to reflect on that level of ignorance in our society and use that as a springboard to ponder the idea of work in a late capitalist economy.
To start with I think this fear of universal unemployment caused by machines was best debunked in the book High Tech, High Touch by the author John and Nana Neisbitt. The authors talk about all of the interpersonal jobs that almost by definition cannot be done by machine. Pretty much any job that requires intimate communication between a client and a worker, think massage or psychotherapy, cannot be replaced by machines, at least not for a very long time, since the degree of empathy and understanding of subjective experience are far beyond where we see AI going over the next generation. Any projection of the economy beyond a generation is beyond us to predict with any degree of accuracy. It is best to limit fear, uncertainty and doubt to something that we can see on the horizon,not science fiction and fantasy.
Of course there is also the obvious point that is frequently made in print and in the help wanted ads, that the talent to improve these technologies is in short supply. The idea that some AI will be able to continually make improvements to its own computer code is purely science fiction.
This is not to suggest that the labor markets are unaffected by these leaps in technology. Obviously they are. And in some cases unemployment is directly proportional to the automation that is put in place. But this is nothing new and has been happening since the beginning of the industrial revolution. Things change and there is pain but we have never run out of jobs. The point we must focus on is not the number of jobs but the skills they need and the processes that change because of the automation.
Some elements within the US culture seem to have a very negative view of higher education. Yet the skill set for the jobs created by AI has grown along with the number of positions. They are so highly compensated that the economic incentives exist. So far they have not been able to fill the demand for the STEM PhD graduates they seek. It is absurd to think that we will every graduate enough PhDs to fulfill what employers currently want. It is an aberation that will correct itself in time. But the larger field of computer related jobs does not shink over time. At most it ebbs and flows and the skills continually shift.
One of the primary areas of focus for those jobs is the automation of human labor or the creation of new information products. Together with robotics it is absolutely true that manufacturing automation has dramatically changed the nature of many industries. The assembly line worker employed on the old Henry Ford assembly line is all but gone. We now have teams of people who maintain the robots, program them, perform quality control checks and many other functions that cannot be replaced. This innovation continues as the cost of these robot machines drops and their capabilities improve.
Organizations are driven by economics. Unless we are to question the very foundation of a capitalist system, that will not change. Using scarce capital efficiently will always be an imperative. Other aspects of capitalism may change but not that. So some jobs cannot be done more efficiently than by humans. Many menial tasks can be changed or supplemented by automated labor but not replaced. Garbage pickup, janitorial service, housekeeping and many other jobs see innovations that use some automation but have never succeeded in completely replacing the human element. And it must be noted that the automation almost always requires additional training for those staff.
In between the two extremes of job categories there are many jobs that appear largely safe even while technology changes the skills required. A modern carpeter, electrician, or plumber uses completely different tools and materials than their grandfathers did. And they must gain the proficiency with those tools and materials to reach maximum productivity. These semi-skilled jobs are still relatively highly compensated with many opportunities for young people. It is social attitudes toward these jobs that often discourage their pursuit.
Doctors had once been more highly compensated than they are today, especially when normalized against the specialized knowledge they must have. This is due to the nature of the profession and its institutions today. Doctors were once solo practioners or in small partnership. They practiced largely unsupervised except for whatever oversight government provided. But the economics of health care has changed that. Most doctors today are employees of some corporation subject to their supervision and direction. It is the corporation which decides the procedures that must be followed and secures the expensive equipment needed for the practice.
The growth of health insurance has also dramatically altered the practice of medicine as well. To pursue increased profitability they will review treatment plans and drugs. They will question or refuse payment when they believe there is a more cost effective option available. Since health care without some form of insurance is impractical today, insurers become another form of oversight and control over the doctors.
These large corporate interests have always sought ways to perform their functions more efficiently and with less liability. Record keeping and standardized processes of medicine may not look like the Henry Ford model but they share the same goal of producing outcomes predictably high quality at low cost compared to a few generations back. They have also offered new ways of doing the same thing better through specialists and machines that can help interpret complex tests. Perhaps the biggest change was the shifting of responsibility from the individual doctor to the institution that employs them.
Many professions have been affected in similar ways by the encroachment of automation and interconnected systems. In the rush to implement these systems I suppose we must excuse some the chaos they create but the implementation of a new system in an office or store has never been painless. This is the point in the economy where I found employment.
Creating computer systems has never been a science. In sofware engineering we talk about methodology. We, like the organizations that employ us, continually seek to reduce risk, improve quality, and do it more economically. As long as we were dealing with systems that could not have life or health impact, we got away with many quality control problems. The money to be saved through a successful implementation was so enormous that even the poor track record of failed projects did not deter the spigot of money released to do new systems. And we did learn how to do things better. But the one common factor here is the management desire to move away from the craftsman error of the pre-Henry Ford model of production to some reproducible and measurable process that could support the continuous improvement paradigm of new management.
Through the decades managements have been partly successful in reducing complex tasks that required highly skilled prople into repetitive tasks that could be done by less skilled people when given very prescriptive instruction. For an example at the extreme let’s take a fry cook at a fast food restaurant. Once the cooks who worked at diners could be highly variable in their skills. Fast food restaurants had the imperative of producing a consistent product but also at the lowest possible price. Where they could they supplemented the job with automation along with training. The job became both more specialized but also more mechanical. Individual variation was squeezed out. The nature of the job itself was changed. Some will say the instrinsic rewards of the job were removed. But clearly the opportunity for individual expression or creativity were eliminated.
Another change enabled by technology is at the checkout counter of a large store. Warehouse stores have learned how to keep shelves stocked with few people and better inventory control due to better machines and software. But the cashier function did not change much until self-checkout. That chapter of American commerce is still being written as the various experiments in self-checkout have found variations on success. Yet the larger picture is clear. The stores reduce their labor expense by asking the customer to do what they once did. Labor is shifting from the store to the customer.
This is also seen so clearly in the embrace of web-enables applications which are substituting for employees, whether at the counter of a bank or on the phone. The rush to push the labor onto the customer has resulted in a failure of training of those employees left to directly talk to customers to even understand their own systems, no less fix what may have gone wrong. So the number of employees may have gone done, but the skill level has gone up. And some companies are inept at training their employees for the new job reducing the overall customer experience.
My larger point is clear. Companies will pursue greater profit, as it their right and duty to the shareholders. But their relationship with their customers is becoming more distant. Most accept this in exchange for the cost saving they may offer. But with consolidation and monopolization consumer choice is shrinking distorting the market feedback to the companies. This may lead to some reckoning in the future…or not. But even this is not my ultimate point.
Automation has dramatically reduced the costs for a lifestyle that would have seemed fantastical some years ago; work from home, home deliver with a day for an enourmous catalog of goods, automobiles with futuristic features. But in exchange something has been lost. We have bought into a narrative that cost is all that matters. True, some have written about this, consider the old saw, knowing the price of everything and the value of nothing. Yet even this I can accept as a price of freedom and progress provided that the ability to choose remains intact. And this brings me to the point I want to make.
The mantra of the libertarian right has always been small government and limited government regulation. There are so many evils papered over in this bumper sticker and I don’t want to unpack it here. Perhaps the tragectory of the 60s and 70s were headed toward a form of socialism that would have been unsustainable. But we did not get to see that instead. The pendulum swung toward the 19c form of laissez faire capitalism. The result of that has been an enormous growth in returns to capital as seen in the stock market but flat earnings growth for the least powerful and skilled workers. Consider what the combined effect of workers who become obsolte for the jobs they have coupled with the lack of retraining opportunities and the high cost of education to enter the workplace and you can see the current predicament of the US society. Now THIS is something I think can but blamed on AI and its associated technolgy and management innovation.
AI was the means by which we made this change to the economy but obviously it was not AI itself but the context in which it was used. In most cases the only reason for automation was to increase profitability even while it was changing jobs which resulted in job loss. It was management choices which did not retrain or replace displaced workers. It was management choices which chose to downsize staffs and ask existing workers to do more with the additional automated support. It was managers who chose to make decisions higher in the management structure and insulate themselves from customers and line level employees substituting absurd versions of the old suggestion box. It was management decisions which only has the singular objective of increasing share price above all other objectives even while corporations merged and consolidate to increase their lobbying and monopoly power. Some suggest this is laissez faire capitalism. Perhaps. But even libertarians do not hold this as the ideal for an economy. And we had rejected this once before. The pendulum must inevitably swing back.
AI will not take our jobs. At best, it will be managements that make us unemployed by withholding the profit they make for themselves and shareholders and not sharing it equitably with the remaining employees who would use it to buy more of the goods and services our economy offers. Instead AI is only serving to improve the extraction of value from labor while experiementing with how low employment and compensation can go before their profitability goes down. Human resources to them is so much underground material to be mined and discarded once the value is extracted. This may be efficient economically but the derived profit accrues to those who control the forms of labor and management decision making.
If this sounds Marxist, it is leading that way, not that his solution is one I want or endorse. The problems of almost all Marxist though is even less attractive than the picture of our late capitalism I have painted. But as Marx and Pikitty have said, capital,wealth, and power pool in an unregulated laissez faire economy over time. It is not the result of greed or the decisions of some star chamber. It is the rules of the game and the emergent effects of those rules when played out over time. No small group of people can change this. Nor can this change over a short period of time. The political process must sort it out and we can only hope it changes without violence.
For the rare person who has read this far it is now time to explain why I find this at all related to the metaverse. You see in its broadest sense a metaverse is a created universe. It is a chance for a do-over or the creation of some utopia. The odds are against it but it is my hope that I can learn the leavers of pressure over the creative processes which could increase the chances of a utopian metaverse over a dystopian one. Frankly I am pessimistic. But I find it a worthy goal and one I’ll continue pursuing. And of course everything I said above could become true of any metaverse that emerges from this orgy of spending on AI and metaverse development. But one can hope for something better.
Post Script
After posting I came across the Oxford Internet Institute and their work on this same topic. I did not read it prior to my post so I’ll be curious to see how my view matches their.
Dale Fletter Blog Site 