Synopsis: AI robots could take us past the limits to new jobs.
It feels like people have “forever” been crying out that new technology would lead to massive unemployment that has never arrived. Like the boy who cried wolf, neither false alarms alone nor past history are a guarantee the threat will never be real.
Everything has limits, and determining the limits requires not just observing historically automation leading to new jobs being created has happened that new jobs emerge but understanding how it happened, and what still applies today. Automation increases productivity, providing remaining workers with extra income and demand for new products and services. This provides scope for new products and services that can employ otherwise “spare” workers. However, there is a limit to viable jobs from new products and services and to such jobs providing for equitable wealth distribution, and even a limit to increases in productivity from automation. While there is uncertainty on when those limits will be reached, it is still prudent to begin planning how to deal that reality.
Even without new job creation reaching limits, there is the challenge of keeping pace with the rate of new technology. Looking at this world economic forum report from 2016, many of the industries at that time suggested as new job growth areas are now on the list jobs threated with redundancies. There were not alone. When this page first published in July 2016, jobs considered safe at that time are now on the list of those jobs threatened:
High end skills, including those of doctors and attorneys, are unlikely to be developed by machines in the foreseeable future and those at the low end such as food service workers are paid such low wages that automation would prove more costly.
RT Feb 2016: Rise of the machines: Robots will replace half of global workforce — RT World News
While back in 2016 the prediction was half the workforce could lose their jobs, given doctors and attorneys are now lists where jobs will be lost and food services workers are already being replaced by special purpose robots, it now in 2024 seems far more than 50% of will lose their jobs, and a lot faster than earlier predicted:
Tesla owner Elon Musk agreed with the founder of the Midjourney artificial intelligence research laboratory, David Holz, who stated that by the 2040s there could be expected to be “1 billion humanoid robots on Earth.”
Jan 2024: Musk predicts that by 2040 there will be one billion humanoid robots on the market
When all is considered, while it is still possible that near full employment could continue and ensure equitable wealth distribution in the coming decades, not having a backup plan would have to be extremely risky.
Technology has always been making jobs redundant: What’s new?
Traditional thinking: history shows there have always been new jobs.
However, despite the predictions of doom, history offers reasons to be optimistic about AI and its impact on work and employment. Jobs have changed and evolved throughout history, which has resulted in the creation of new professions that were previously inconceivable. For most of the 20th century, typing was seen as a desired and decent job, and typists were in high demand.
AI will take some jobs, but mass unemployment isn’t inevitable
As computers grew in popularity and typing got easier, the demand fell away, and the profession nearly became extinct. But, thanks to the same trends, the demand for web designers, graphic designers and copy editors increased.
This is the traditional approach: new occupations have always followed the loss of jobs to technology in the past, so this will continue to happen in the future. However, the example of computers leading to end of the job “typist” and the “typist pool” was not as simply as new technology replacing the typing pool, and more about a change of culture as everyone learned to use computers. The more typical example is something like devices that directly make people redundant:
Elizabeth I, who refused to patent a knitting machine invented by William Lee, saying “Consider thou what the invention could do to my poor subjects. It would assuredly bring them to ruin by depriving them of employment, thus making them beggars.”
Wikipedia: Technological unemployment.
But the typing example does highlight that there are lots of ways jobs become redundant, and the point about history is so far correct and new jobs have always emerged, although, unlike as proposed in the Guardian story, not because the new technology directly created those new jobs. If new technology did always directly create new jobs, then the technology would not really reduce required labour at all, so there would never be workers available to work in entirely new jobs and new industries.
But “have always” is not quite the same as “will always”. Nothing keeps going forever, and things only can keep going until a limit is reached. Picture a person in a phone booth who invites in one extra person. They fit! Then they both invite another, and they again fit! But this would not prove that people can just keep being added indefinitely. Then there is the story of the man who fell from the top of a tall building declaring as he passes each floor “so far, so good!” Everything reaches a limit.
A key question becomes: “Can we ignore that there must be a limit to possible new jobs, if we have not yet seen evidence of the limit?”
How and why have there always been new jobs?
Image a small, self-sufficient country that has economy producing exactly sufficient for all the needs of the population, and in this oversimplified society, everyone earns exactly what they need to in order purchase exactly what they need.
But what happens if in this society, a significant percentage of the people had been employed making clothes until most of them are made redundant by new technology.
The work previously done by these clothes makers is now done without them, which means the country can still provide for the needs of the entire population, including the clothes. In fact, the clothes factory can now make more profit on the clothes it sells since the cost of making clothes has dropped. The clothes factory could pass on some of the savings to customers, reducing the cost of clothes, and thus the cost of living for everyone! Between the increased profit for the clothes factory and the lower cost of living, everyone can have a little extra available to spend!
Well almost everyone. What about those who were made redundant? Until they find new work, they are now a wasted resource, and the loss of their consumer spending will mean some production now goes to waste and all businesses suffer from the reduced consumer spending. The result could be an economic disaster!
Provided all these people have at least one hobby or recreational sport, declare the people now professional at their favourite hobby or sport. Now they are again being paid they can restore consumer spending and ensure everyone has their previous incomes, and the rest of the population can spend their savings from the reduced cost of living on the products produced by the hobbies or by watching the new professional sports people. Increases in productivity are what enable activities like playing sport to become professions.
The reality is it does not matter what work is found for those who became unemployed, as anything that is even marginally productive will add to the economy, and all maintaining consumer spending which is the main metric for the economy.
This illustrates how, provided the new technology increases productivity, then the introduction of the new technology provides funding to employ workers made redundant in careers that could not previously be funded by society.
While new careers, like web designers etc, may arise, the real reason improvements in productivity continues to create jobs from the increase in productivity which provides the money for people to consumer ever more goods and services.
Concluding that “the stock of work in the economy is not fixed”, the report cites the surge in hairdressers as evidence that where one avenue closes in the jobs market, others open.
The Deloitte economists believe that rising incomes have allowed consumers to spend more on personal services, such as grooming. That in turn has driven employment of hairdressers.
So while in 1871, there was one hairdresser or barber for every 1,793 citizens of England and Wales; today there is one for every 287 people.
Technology has created more jobs than it has destroyed, says 140 years of data | Economics | The Guardian
As the report quoted above concludes: “the stock of work in the economy is not fixed”. However, that does not mean it is unlimited.
What has changed from adding automation in the past?
Confidence that adding automation creates sufficient jobs to offset job losses is based on what has happened in the past. It makes sense to consider what has changed between the past and today, in order to be confident that the same rules still apply.
What has changed?
- There are no ‘new lands’ creating expansion of ‘civilisation’ into new areas. (There is no place to grow a new USA).
- We are no longer living through a population global explosion.
- There is no simple repetition of the emergence of the middle class, as developed countries already have an established “middle class” and economic development in developed countries is no longer providing a huge lift in living standards for such a large portion of the population.
- The capability of gap between automation and humans is being eliminated and AI now has intelligence and can now replace even doctors and lawyers and highly skilled professions.
No new lands: The ‘new world’ has already been populated and explored.
Most technology development happened during the growth age, a time when the societies making the technological advances were going through rapid population growth and were forming empires and expanding into a whole new world.
Although throughout much of the industrial revolution colonising new land played a significant role in economic growth, this was largely offset by global population growth.
Now in a post “peak-child” era there is effectively no ongoing source of global population growth, but many significant economies will experience population growth through immigration for potentially decades to come.
This change will change the nature of employment but is not seen as changing the ability for economic growth per capita to continue to lead to new jobs.
The population explosion is ending.
While the population explosion is ending, the market population is still continuing to grow at a much-reduced rate as more of the world’s population move to living in more developed economies.
Again, this is noted, and will slow economic growth, but not necessarily economic growth per capita or the ability of productivity increases to lead to new jobs.
The middle classes are either no longer growing, or at least growing as fast.
The societies experiencing the change to new technology were also experiencing significant income growth for most of the population and experienced the emergence of the middle class.
The biggest impact of this change is that a lift in economic circumstance is now less linked to social status for a large percentage of the population.
To be explored further but so far appears to change motivation for increased productivity rather than any ability for further increases.
Capability of technology relative to humans.
Technology has traditionally been in the form of tools that empower humans to produce more, rather than full automation that allows production completely without human operators.
Technology has previously had no ‘intelligence’, and has therefore been limited to replacing mundane repetitive tasks, and still needed people to supervise and play a role
The closing of the gap between what can be done by human and machine and what can be done by machine, or at least AI or AI robot alone, does clearly impact the ability to find new jobs for humans.
Limits to increasing productivity of workers.
Productivity it simply the production per worker: productivity = production ÷ workers.
Either increase production or decrease workers to increase productivity.
Limit to the ability to increase production: There has to be a market.
One key factor that has changed from the 20th century is the rate of population growth. During the population explosion, all markets continued to increase in size. This meant new automation could easily increase productivity by allowing a business to simply increase production to supply an ever-increasing market.
With the flexibility of AI robot technology and a stable global population, the alternative approach of increasing production by increasing the variety of products produced, but this not only increases productivity, it also absorbs the job creation that would otherwise arise from the increased productivity.
Limits to the ability to reduce labour.
Some things can be reduced to just one worker, others to even zero workers. Either way, once you reach the one or zero, you can’t go any further. The closer to that one or zero limit, the less the impact on the workforce that becomes available to have an impact on creating productivity in a new industry.
Limits to the creation of new jobs.
The limitation of shifting the sector providing employment.
For the earliest humans, the only occupation could have been obtaining food. Housing, toolmaking and the production of human accessories in general would have soon followed, with almost all human engaged in what we now call “primary and secondary industries”.
Throughout history, as technology evolves productivity increases resulting in more and more “tertiary industry” jobs.
The introduction of AI Robots could potentially move almost the entire population into the tertiary sector, but after that, where does everyone go next?
Perhaps 100% is the limit to the percentage of the population who can be shifted from primary and secondary industries into the tertiary sector, but even if that is the limit, we have already progressed a long way towards that limit.
Do we need to create a quaternary sector, or have we perhaps even already created one but are not yet tracking it?
Limits to productivity of new jobs.
Douglas Adams in the novel the The Restaurant at the End of the Universe speaks of an “Ark Ship” of Golgafrinchans:
This particular group consists of the Wodehousian[7] “middle class” who have common, middle-management types of occupations. They were sent away from their planet under false pretences by the (upper class) “thinkers” and (working class) “doers” of their society, who deemed them useless. They were told that the entire society had to move to a new planet, with a variety of thin excuses, and that it was necessary for them to go first to prepare the new planet for their occupation.
List of The Hitchhiker’s Guide to the Galaxy characters
The humour is based on the thought that there are now many occupations that may contribute questionably little additional value to society.
The whole benefit of increasing productivity is that it frees up workers to then produce that which no one was previously free to produce, but it this new production is of questionable value, then the increased productivity is providing little benefit.
Since the start of the industrial revolution, the percentage of the population employed as professional sports people has continued to increase, but is there a limit to the percentage of people who can be sports professionals?
The risk is that the more jobs that have already been created, the further down the list of the most useful new things there are to do we progress. Is there a limit to this list?
Limits to what jobs need human rather than robots and AI.
Another limiting factor may be that it only makes sense for to create new jobs for those made redundant by technology, if those new roles are not also best performed by AI and/or robots.
For a job to provide employment under the current system, the job must provide sufficient pay to deliver an acceptable living standard for human being, and the cost of AI equipped robots is lower than the amount needed to provide an acceptable standard of living.
Prior to AI and AI robots, automation was almost always designed to automate either only one task or a very limited range of tasks, and had to be designed, constructed and programmed for those tasks. This meant that there would always be a delay between the new jobs being created and availability of technology to automate those new jobs, but now, with humanoid robots that learn through AI, robots already in production may be able to learn a new job without programming even quicker than a human can. Plus, once one robot has learnt how to complete a new task, their learning can be instantly copied to every similar robot.
Limits to a system reliant on jobs for wealth distribution.
At COP28, wealthy nations committed just over $700 to a “loss and damage fund” to compensate less developed countries impacted by climate change caused by the emissions of those wealthy nations:
Wealthy countries most responsible for the climate emergency have so far pledged a combined total of just over $700m (£556m) to the loss and damage fund – the equivalent of less than 0.2% of the irreversible economic and non-economic losses developing countries are facing from global heating every year.
$700m pledged to loss and damage fund at Cop28 covers less than 0.2% needed
This fund is intended to compensate thousands if not millions who have lost their lifetime livelihoods across many nations, but perhaps interestingly, the total amount committed, curiously the same total as the 10 year contract just signed by baseball player Shohei Ohtani:
Shohei Ohtani — deferring $680 million of his contract for 10 years — may face some financial risks, advisors say
- Shohei Ohtani made history this week with a 10-year, $700 million contract to play for the Los Angeles Dodgers.
- The Japanese superstar will only receive $2 million per year over the agreement and will defer $680 million.
- Some of the risks of deferred compensation include higher federal taxes, inflation and company solvency, experts say.
It will take the baseball player 10 years to earn that amount money, while head of UK based Bet365 can earn that amount in just over 1 year, and Taylor Swift can earn over $1 billion dollars from one tour.
While for those being compensated for lifetime earnings, their share of the $700 million is, at this time, all they will ever receive as compensation.
While it is unclear what someone enabling gambling would have earned in the past, the rise in the pay of highly paid musicians and sports stars may have increased beyond their increased contribution to total productivity over and above other contemporary and historical musicians and sport stars who receive far, far less.
Overall, it seems very possible that the current system of wealth distribution has some serious issues, and that they are getting worse.
Solutions providing alternatives to near full employment?
Tech billionaire Elon Musk has predicted that artificial intelligence will eventually mean that no one will have to work.
He was speaking to Prime Minister Rishi Sunak during an unusual “in conversation” event at the end of this week’s summit on AI.
The 50-minute interview included a prediction by Mr Musk that the tech will make paid work redundant.
He also warned of humanoid robots that “can chase you anywhere”.
The pair talked about how London was a leading hub for the AI industry and how the technology could transform learning.
BBC November 2023: Elon Musk tells Rishi Sunak AI will put an end to work
Technology means the work will still be done and the wealth still created. But we don’t only rely on jobs for the productivity and wealth creation, but also for wealth distribution and personal fulfilment.
Even if society can produce everything required for a highly desirable standard of living for the entire population without anyone at all doing any work, then the two problems of how to distribute the wealth and how to gain personal fulfillment would still remain.
Since it is difficult to imagine how everything could be achieved with planning, philosophy and even just people to continually revise the goals of society, it is probably safe to assume there will at the very least still be come jobs in the future, and thus some possibly and incentive in terms of distribution for those who make a greater contribution to society.
Suggestions such as a basic income help with wealth distribution but potentially create significant social problems from lack of self-esteem. The major problem to these schemes is the free trade pressures may simply make such measures impossible.
Solutions that based on stopping the development of technology involve more troubling isolationism but need further exploration at some time.
Reports & References:
Reports on how jobs will be lost to robots even predate the recent emergence of generative AI and the humanoid robots that only started to become reality in 2023.
2016 Reports.
A NEW study has found that almost 50 per cent of the work force will be replaced by robots within the next 20 years. Find out which jobs will be the first to go.
SCIENCE fiction writers have long anticipated technology taking over the workforce but it seems the reality is close upon us.
A new study has found that 47 per cent of jobs in the US are “at risk” of being automated within the next 20 years.
The researchers analysed more than 700 jobs listed on a careers website along with the skills and education required for each position and weighed them against how easily they could be automated and what engineering obstacles were preventing them from being computerised.
Specifically the researchers say that “low-skill and low-wage jobs” will be the first to go.
“Our model predicts that most workers in transportation and logistics occupations, together with the bulk of office and administrative support workers, and labour in production occupations, are at risk,” the researchers wrote in the Oxford University study titled “The Future of Employment: How susceptible are jobs to computerisation.[link dead]”
News.com.au: Robots Take Over the Workforce.
In the 1800s it was the Luddites smashing weaving machines. These days retail staff worry about automatic checkouts. Sooner or later taxi drivers will be fretting over self-driving cars.The battle between man and machines goes back centuries. Are they taking our jobs? Or are they merely easing our workload?
A study by economists at the consultancy Deloitte seeks to shed new light on the relationship between jobs and the rise of technology by trawling through census data for England and Wales going back to 1871.
Technology has created more jobs than it has destroyed, says 140 years of data | Economics | The Guardian
Will robots take over the workforce? And if—or when—they do, what jobs will be left for us humans?
Wall Street Journal March 2014: SXSW, For One, Welcomes the Rising Robot Workforce – WSJ
Updates:
- 2024 March 28 : Second edition. Most content new, but some only significant updated.
- 2016 July 11 : First edition published.