Automation and society

Automation – the end of all (our) work?!

Experts predict digitization and automation to take over a third of our work by 2025. Have you already planned your holidays? Maybe you should keep a flexible schedule…

Visions of an automated world have been around since the start of industrialization. Will new technologies be the catalysts for human redundancy? Where we are heading?

A reoccurring theme in history

Already back in 1772, the British writer Thomas Mortimer stated in his “Lectures on the elements of commerce, politics, and finances” that “those [machines] which are intended almost totally to exclude the labor of the human race.” and that machines “if introduced into our dockyards etc. would exclude the labor of thousands of useful workmen.” And an article from 1955 in the Time Magazine took an optimistic approach by explaining that the “automation of industry will mean new reaches of leisure, new wealth, new dignity for the laboring man.”

Do these statements sound familiar?

These kinds of visions of the future seem to be a reoccurring theme throughout history. Nowadays – just as in 1772 and 1955 – we again read about machines expected to take over a huge amount of jobs. Here are two examples:

Once a door closes, another one opens…

It seems these visions of a world with high unemployment, or in a more flowery description, of personal lives with an abundance of leisure time, have always been based on the structure of the job market at that very time. Still, for centuries now, technological advances have indeed killed a majority of jobs and revolutionized the job market over and over again.

In 1870, for example, almost 50% of the US population was employed in agriculture. As of 2008, less than 2% of the US population is directly employed in this area – you can find more details here. Affected farm workers, however, have moved forward and found new jobs in production factories or mines.

On the one hand, these jobs slowly but surely have been rendered redundant by (more sophisticated) machines or the emergence of new (mass) production technologies or materials (e.g. plastics instead of wood), and on the other hand, developments have given rise to a service society that created a completely new set of jobs.

So while initially, unemployment rates have always risen whenever new technologies have been introduced, this tendency has eventually always been made up for after a while by the emergence of new job opportunities.

In short, new technologies have always been the door-openers to new products and services.

Today’s vision of automation

Until the 1970s/1980s, automation mostly referred to mechanics. Since then, the rise of IT continuously added new means to automation, which today, culminates in the encompassing vision of a digitized world.

I, personally, have no doubts that the digitization trend will establish itself and further automation of a lot of jobs will indeed come to pass. It’s just a question of how fast and how far. The jobs, or let’s say the “tasks”, that we are looking to automate are increasingly complex and as a result, so are the machines that are to assume them.

I very well remember my visit to a Volkswagen manufacturing plant in the early 90s. What I saw there was both a nearly deserted chassis production line as well as some disillusioned managers stating that it had been over-automated. At the time, the trend was CIM (Computer Integrated Manufacturing). The manufacturing robots proved too unreliable and required significant servicing efforts so that downtimes and maintenance resulted in higher costs rather than the anticipated savings.

Admittedly, by applying this 25 years old experience of mine to today, I might be comparing apples to oranges. But I think it is still a good example to highlight a few questions and road blocks that may influence the pace of digitization advancement, especially when the automation involves the physical world.

Automation – the road blocks

Undoubtedly, automation replacing human labor could be an answer to various challenges. Whether it is to fight ever-increasing labor costs on a global level or to replace the retiring baby-boom generation that will be succeeded by a much smaller workforce in some western countries.

This is further fueled by decreasing costs of machines that drive automation and their steadily increasing output. The cost of owning and operating a robotic spot welder, for instance, has tumbled from $182,000 in 2005 to $133,000 last year, and will drop to $103,000 by 2025 – as estimated by Boston Consulting.

What are the hurdles though? Here are a few examples of what we need to consider:

Legal framework

Governments (and society) are overwhelmed by the speed of technological opportunities arising and struggle to provide required legal frameworks on time. This hinders their use as we can see in the areas of autonomous driving or drones, for example.

Downtimes/maintenance

Ever-growing complexity of tasks to be automated leads to machines becoming equally increasingly complicated. Complication is the natural enemy of reliability. Predictive maintenance, for example, may be a solution to reduce unexpected downtimes but it will not lower the high maintenance costs.

Robots-human-incompatibility

How long will it take mankind to accept the interaction with robots? Despite some positive examples, people generally have a hard time accepting robots in areas in which they were previously used to human service. So, what will happen first, robots becoming more human or humans becoming more machine-compatible?

Controlling the automated world

In order to control and manage an automated real world, a digital twin of it is required (unless we talk about fully decentralized and autonomous decision-making). Can the digital copy of a complex and changing real world be developed and frequently adapted fast enough to avoid a loss of control?

Sensor dependency: sensors are very critical elements of future automation. Only they provide the increased and required degree of flexibility – compared to old machines that could only operate in limited, predictable environments. As such key elements, will sensors be reliable enough?

Demand-fluctuation

Machines cannot engage in part-time work and do not have a flextime wage record to cope with economic downturns. Nor can machines be transferred to other job assignments as easily as humans. And machines are subject to fix costs to a large extent – regardless of the level of deployment. You might say that machines will not be purchased anymore in the future but merely their services? Yes, that’s fine, but somebody will have to pay the bill at the end of the day and charge it back somewhere – sooner or later.

IT security

An automated world is highly sensitive to any disruptions in the IT-layer. This includes network failures and hacker attacks. Will we be able to handle the risks?

Ethics

With machines taking over more and more decisions that have been made by humans so far and, with that, also increasing their level of responsibility, we will face truly tough discussions around ethics. It’s not a question of whether machines or humans are doing the better job, but a matter of rule sets for machines that we humans need to carefully consider and consciously decide upon.

Conclusion (for today)

Regardless of the pace of automation: the established cycle of people losing their jobs in light of new developments in the past will continue today – and tomorrow. In the past, however, developments did not advance as fast as they do today, and the efforts involved in finding new jobs and/or acquiring new job skills nowadays (or in the future) are significantly higher.

The alternative – changing to a profession that requires less skills – is not really an attractive option, because such moves typically involve job roles that are also likely to be automated in the future. Re-adjusting educational systems and helping people to more flexibly navigate through an evolving job market is a real challenge for society.

As mentioned before – when one door closes, another one opens. Do you find yourself struggling to imagine what such new job opportunities could look like in the future? My tip: just set your mind back in time by 15 years and then check out these two links:

I’m interested in hearing your view on this and look forward to your comments on LinkedIn.