April 11, 2017    8 minute read

How to Stop Automation and AI from Worsening the Ageing Population Problem

Time for a Robot Tax?    April 11, 2017    8 minute read

How to Stop Automation and AI from Worsening the Ageing Population Problem

Millions of young workers out on the streets protesting against robots, calling for the abolishment of AI – does that seem a likely outcome in 10-15 years’ time? It will be the case if action is not taken now, alongside an improvement in how to understand AI’s benefit to the workplace. This article sets out the scene on where societies are headed, and how AI will play a role in the transition to the new world.

After the Berlin Wall was constructed, West Germany faced a labour shortage due to the restricted flow of labour. As a result, back in the ‘60s, West Germany invited workers from Turkey, among other countries, to support the backbone of their industries – many of who ended up staying there.

AI and an Ageing Population: Recipe for Disaster?

Now, many question if it will be the same case for the Western nations with their own ageing population problems – whether, one day, they will be forced to proactively encourage foreign workers into their country, when no other labour source is doing the job. With the ageing population that is visible from the shift in the population pyramid (below), a smaller proportion of the younger end of the population will have to carry the burden of the ever expanding elderly population among some Western nations.

Ageing Population Pyramid

Source: PopulationPyramid.net

Inviting foreign workers could well exacerbate existing cultural tensions, giving more power to populist and often protectionist politicians, threatening to add more weight to the economic burden. But it’s important not to forget how changing productivity could alter this trend, with one major aspect of the future workforce being the power of automation through AI.

Interactive Robots Lend a Hand

Until recently, many dismissed such power because robots lacked the initiative their human counterparts can take. They were programmed to do tasks faster and more efficiently, but they could not decide under unknown or unpredictable circumstances. Today, robots are no longer blind and deaf. They are being trained through ‘machine learning’ to observe and react to their surroundings with the assistance of sensors and cameras, for example, and their abilities can be enhanced with temperature sensors and other additions. These developments will make robots more aware, increasingly reactive and responsive to events even humans cannot predict.

AI algorithms will be able to do specialised tasks much more efficiently and with lower rates of failure, such as the robot in a hospital pharmacy handing out 6 million prescriptions with an error rate of 0%, compared to 2.8% when it’s humans delivering the medicine. With such improvements, large parts of the world’s youthful population will be rendered economically and physically redundant, causing countries with a predominantly young and undereducated population to suffer the burden, whereas Western countries will transition smoothly into the new world they have created.

When One Human Counts for Less

Most of the proponents of the original idea that developed countries will need an increased number of younger workers to support their economies have arguably not paid enough attention to 2017. Or, to put it more conservatively, they could not have predicted that it would be happening so soon.

Yuval Noah Harari, whose book Homo Deus examines the possible future paths for humanity, argues that this will have even more serious consequences for the social structures that define the human role in society. Up until the last few decades, almost every single human being was important, whether it was for war, industry, agriculture or any other area. For that reason, the political classes invested in areas such as public healthcare and education, in no small part due to their goal of improving their countries’ economic power. That appears to be changing in the modern era.

Is AI Feeding Income Inequality?

This division is readily visible in the USA, with income inequality nearly at its peak. The elite no longer need to care about the masses, since the latter have diminishing economic value. Naturally, the young and undereducated masses will not have the skill sets the top tier will be looking for.

Source: Inequality.org

With rising unemployment in the mid-to-higher income groups, the effects to the economy will be detrimental on an aggregate scale. When the income group that buys the products of those automated companies is stripped of its income, demand will fall severely, risking an economic recession.

Slowing Down

Given the real possibility of mass automation in man sectors of the economy, it is perhaps easier to grasp the idea that countries with ageing populations will not need the projected number of young workers. This, however, brings up a new problem. Although it looks likely that demand for labour will be pushed down by automation, this does not mean that companies will pay remaining employees more. It is also expected that changing long-term market fundamentals mean that people at the age of 30 “today will have to work seven years longer or save almost twice as much to end up with the same nest egg as those of roughly a generation ago”, underscoring the upcoming squeeze on the millennial generation.

The Rise of the Robot Tax

There is one way to prevent this problem – something that most companies would not want to hear. To sustain the pension system at reasonable levels without putting too much load on future generations, governments can start taxing automation. Although this will slow down the shift to automation and cause inefficiencies, it could alleviate the long-term burden that millennials will have to face in the future. This short-term strain on companies to extend the period of technological and economic growth will secure a better future for western nations’ societies.

There are many possible ways to tax automation, and one of them is as follows. Every robot has a labour equivalence (i.e. a company can lay off X number of employees per robot). If this can be tracked or demanded by audit requirements, then governments may issue an automation taxation calculated as a percentage of the potential benefits that automation is expected to generate. The funds from this tax inflow could be funnelled towards a centralised pension fund, or the company’s own pension fund.

This taxation system would also scrutinise companies over their investments in automation. The cost efficiency from automation would have to create value well above just laying off workers in order to compensate the burden from the additional taxation. As a result, companies would not switch to automated production for the sake of switching, and only do so when it is significantly effective. They would also have to increase R&D funding in order to create better solutions, strengthening technological advance.

Foundations for the Future

This taxation structure would not only serve to mitigate the potential rise in unemployment in the near future, but also to lay the grounds for the upcoming AI revolution. Most debates focus on robotics and unemployment in labour-intensive sectors such as agriculture or manufacturing. Those are the short-term concerns regarding unemployment. However, when the topic is population structure, the time horizon is not 5 to 10 years, but 15 to 30 years.

Given this, even though AI is not overwhelmingly smart today, it is important to consider what it will be in 2030. Flowing into reasonably secure sectors such as education, AI could displace workers more than is often anticipated. For example, in 2016, the Georgia Institute of Technology in Atlanta used an AI-based teaching assistant to aid in running an online course – yet almost none of the students realised that their questions were being answered, and discussions being encouraged by, a computer programme.

Looking Forward Without Fearing It

Some oppose the view that AI will take over highly-skilled jobs, as machines cannot feel what humans do. But what if they do not need to feel in order to replicate everything a human being can do? After all, everything a human being feels is merely an output of a chemical algorithm, and any algorithm can be replicated. At that point, denying AI the chance to become reality out of fear does not change the facts, and resistance to it is futile.

What can be done instead is to set a framework for current unemployment concerns that will be the foundations on which the industry can work on in the future. AI is not out to destroy humanity, but humans still need to decide how they will live with it.


Get articles like this straight to your inbox each morning with our Breakfast Briefing. Sign up by clicking here!

Log in with your details


Forgot your details?

Create Account

Send this to a friend