Will automation lead to an economic crisis?

I’m probably not the first to write on this topic. But I really want to know the opinion of others. I am not an economist or an expert on the subject. This is my attempt to find out what economists and experts think so that I can understand the problem.

In my opinion, until the pace of automation grows faster than new jobs can be created, AI should not be expected to cause mass unemployment or anything like that. When AI can find a new job as quickly and cheaply as a human, then the economy will collapse (but everything else will collapse too, because it will be the Singularity).

As software and hardware become more powerful every year, many are worried that workplace automation will lead to some kind of economic crisis. This can take the form of persistently high unemployment, lowering the wages of many workers below the subsistence level, or abrupt transitions to another economic system in response to these conditions.

If you imagine Asimov-style androids with positronic brains, it’s easier to imagine a world in which all jobs are automated. In reality, though, it would be a waste of resources if robots literally came in and did the job as temporary replacements for workers, and there are a few professions where that would make sense. Many programs in the future will be more human-like in the sense that many machines will be able to have natural language and image understanding capabilities, as well as be able to reason about the larger context in which their work exists, to avoid dangerous or costly errors due to to a lack of common sense. However, in many other respects, the software for almost all working robots will be very different from the human mind.

Some jobs, such as picking most fruits and vegetables or most assembly line work, cannot currently be done by machines as they require manual dexterity. In such cases, you will need more precise manipulators with touch sensors that can adapt to a wide range of situations. However, there is no reason to expect that they will look like human hands. Of course, machines will also be networked, so it is also unrealistic to imagine them as a bunch of people. Finally, the way tasks are currently divided into jobs makes sense for human workers, but doesn’t make sense for automation. Instead, some tasks will be automated first, while the rest of the tasks that are part of the job will still be done by humans.

What is happening with technological unemployment today?

Worldwide, employment rates are not worse than historical levels. This seems to show that jobs are created more or less as quickly as they are automated. Some of the jobs that are expected to grow are in healthcare, renewable energy and a few computer jobs. Among those who are shrinking are secretaries, writers, assemblers of equipment. In recent times, automation has tended to take over semi-skilled jobs.

Another metric to consider is GDP. Before 1000 AD GDP per capita was below $1,000 adjusted for inflation everywhere. In Western countries today it is about 50,000 dollars. Since human innate abilities have not changed, this must be the result of innovations (in education, processes, tools, equipment, etc.) that allow people to produce more value for every hour of work. On the one hand, this means that there are already 49 automation “robots” for every person in these countries. Nevertheless, employment remains at about the same level as always.

So we know from experience that employment can adapt to extensive automation. However, these changes took place over millennia, centuries and decades, respectively. If the rate of job automation were to increase so that a large proportion of jobs were replaced over many years rather than decades, people might not have time to retrain fast enough to avoid some higher unemployment. However, again, this does not seem to be the case at the moment.

Are there things computers will never be good at?

The new methods that the researchers have come up with are not only able to perform better than previous methods on a particular test; they are also becoming more common. Artificial general intelligence is what is commonly referred to as humanoid AI these days because the ability to deal with unexpected situations is a central part of what makes human intelligence special.

What jobs will be automated?

These will most likely be jobs that: do not require fine manual dexterity, originality, or people skills, and are therefore likely to be automated in the next few decades. This includes most office and administrative support jobs, sales jobs, some service jobs, and most manufacturing and transportation jobs.

In addition to inventing hardware and software capable of performing these tasks, the cost of developing and deploying the technology must be less than the cost of hiring workers so that they can be replaced. The cost of computers has been steadily falling for several decades, and there are no fundamental physical limits to this improvement that would prevent the trend from continuing to the size and power consumption levels of the human brain. Robotic bodies and manipulators, while their dexterity, sensory abilities, and cost continue to improve over time, are unlikely to have the kind of exponential improvement we see in computing hardware. Again, we know that a machine with human dexterity is possible (because hands exist), so it seems inevitable that machines will eventually surpass us in these abilities as well.

However, there are certain jobs that some people are willing to continue paying a human to do, even if the robot does it better in some way. For example:

production of handicrafts
creation of works of art whose value depends on whether they are originals
some types of cooking
performances (acting, dance, comedy, etc.)
housework (personal servants such as gardeners, etc.)
sport
therapy
hairdressing and the like
massage
some aspects of medical care (feeling of care)
some aspects of training (motivation, mentoring)
some aspects of war (decisions about when to use violent force)
work of the clergy
mortuary services
some types of sales
policy

For such tasks, human performance is part of what is appreciated by some customers.

As more jobs are automated, what economic effects should we expect?

Around 1800, the economist Jean-Baptiste Say argued that workers laid off because of new technology would find work elsewhere once the market had time to adjust. By the mid-1800s, there was a theory that explored the economic impact of automation. AT “Capital” Marx would later call this “compensation theory”. This includes additional employment in the capital goods sector, price cuts, new investments and new products.

According to this theory, when workers are fired because their jobs are automated, this frees up capital, which the owner then uses to hire other workers to do other jobs. Therefore, the number of employed workers does not decrease due to automation. (Marx disagreed with this, saying that part of the capital would now be associated with machines). The theory also discussed other effects. Automation reduces the prices of goods, making them more affordable. It also lowers component prices, making new products viable. Companies producing these goods earn more profits, which allows them to expand and hire more workers.

Because of these effects, as long as the market has time to adjust, we should not expect the unemployment rate to rise to the point where robots take over all jobs. Instead, new jobs should be created for workers until the entire employment reserve is used up. This process can be expected to continue until all the work is done by machines at lower cost. As long as there are skills that people can do cheaper than machines, the number of jobs using these skills must increase until they consume the entire available pool of human labor.

The more jobs that are automated, the higher the average standard of living will be, as the amount of value produced per capita will rise higher and higher. Even without higher tax or charitable rates, the total amount received will increase as more goods are produced at less cost. Whether this will result in more people living on unemployment benefits or not is more a matter of political argument than technological extrapolation.

Conclusion

The economic benefits of all this automation will flow primarily to those who own the machines. The more they invest, create new products, spend more, pay more taxes, and civilization as a whole benefits from this, although some people will certainly not be able to adapt and find new work and will suffer. How we decide to provide for those who cannot find work is something that each country will have to decide for itself.