While people have frequently feared that automation would lead to mass unemployment, it has never actually happened.
Image: Eric Limon/Shutterstock

The grids are all right: Robots have never and will never steal jobs

9 Oct 2017

They probably aren’t going to allow you to live in Oscar Wilde’s work-free, creative, socialist utopia either.

Future of Work Week

On a November night in 1811, angry workers smashed textile machines in Nottingham. This wasn’t the first time something like this had happened – the first such incident, on 11 March of the same year, had to be disbanded by British troops. No one managed to catch a glimpse of their elusive leader, Ned Ludd.

Unbeknownst to the protesters, they had inadvertently established a long legacy of the term ‘Luddite’ forever being associated with fears that automation would lead to mass job loss. This is despite the fact that there is no evidence that Ned Ludd ever existed – he was a historical figment, a false firebrand, but powerful enough a myth that the term entered the common vernacular and never left.

Lately, livelihood-related anxiety is rampant because the landscape of the workforce is changing so rapidly. Automation has become so advanced that we as a society are now flirting with the concept of AI, may soon have access to self-driving cars and have confirmed reports that there are thousands of robots scuttering around Amazon warehouses collecting packages for dispatch.

The ebb and flow of industry

This new reality, coupled with a recent massive depression in employment levels, means people are beginning to panic that automation will trigger job destruction at such speed that human enterprise will fall totally into obsolescence. Historically, however, there is nothing to support this idea.

In 1900, 41pc of the US workforce was employed in agriculture. By 2000, that figure had fallen to 2pc. Myriad new technologies introduced over the course of those 100 years, such as automated machinery, are primarily behind this decrease in numbers.

Yet, as MIT economist David Autor has pointed out, the employment-to-population ratio rose during the 20th century – despite both technological progress and women newly entering the workforce.

From the cotton gin to steam power to the common calculator, new innovations in technology have been revolutionising industry for hundreds of years. For every leap in progress, some form of industry ceases to exist. The invention of the car, for example, meant that the number of people employed in equestrian-related industries fell.

Long history of Luddite fears

The dearth of evidence in support of automation-related job loss has not prevented people from repeatedly panicking about the possibility. One of the machines often targeted by the Luddites, the stocking machine, was originally denied its patent for this very reason.

Queen Elizabeth I crushed the hopes of inventor William Lee – as Daron Acemoglu and James Robinson explain in their book Why Nations Fall – by turning down the patent, citing her concerns about job loss.

‘Thou aimest high, Master Lee. Consider thou what the invention could do to my poor subjects. It would assuredly bring to them ruin by depriving them of employment, thus making them beggars’

When Lee’s machine eventually received its patent, it didn’t put people on the street. If anything, the increase in productivity led to industry growth, which lead to more job creation. This tends to be the pattern over the years of technological progress.

The trope of robots leading to humanity becoming totally irrelevant has endured in popular culture. Kurt Vonnegut’s 1952 novel Player Piano spoke to concerns over excessive automation.

The Twilight Zone crystallised these fears in its 1964 episode ‘The Brain Center at Whipple’s’, detailing a dystopian world in which all workers have been replaced with robots.

Gimme shelter (from the robots)

The ’60s was a notable flashpoint of Luddite fears. Automation anxiety built up so significantly from the ’50s into the next decade that, by 1964, US president Lyndon B Johnson created a Blue Ribbon National Commission on Technology, Automation and Economic Progress.

The aim of this panel was to assess the risk that productivity would rise so fast that it would outpace job creation and lead to a reduced demand for labour. The commission concluded that automation did not threaten employment.

This did not stop the commission from recommending, as Autor demonstrates using excerpts from a 1966 edition of The Herald Post, a “guaranteed minimum income for each family … two years of free education in either community or vocational colleges [and] a fully administrated federal employment service”.

A robot will never ask if you want fries with that

A future of restaurants filled with meal-dispensing machines and automated waiters is a distant reality. The image of the sound of a fork-drop reverberating loudly through an empty eatery, now devoid of staff, is a gross over-exaggeration. As the technology stands, it is unlikely to ever be feasible.

Jobs such as waitressing, as well as something like hairdressing, require flexibility and visual recognition. Given that Google needed 16,000 computers to allow an AI to recognise cats in video, and given that robots were still struggling as of 2013 to recognise a chair, the technology certainly isn’t at the point where it could differentiate between a burger with onions and one without.

It is easy to teach robots tasks that are repetitive and predictable. Lack of adaptability and improvisation are currently one of the core weaknesses of existing AIs. It goes without saying, also, that sometimes people actually enjoy the engagement with staff central to getting a haircut or going out for a meal. These industries are not amenable to automation, so jobs in those sectors are secure for the foreseeable future.

The jobs most at risk, somewhat counter-intuitively, are jobs in sales and admin. Bookkeeping, clerical work and certain manufacturing jobs are easy enough to automate. As Autor notes, this process has already begun and, since the 1980s, automation has started to take over these tasks.

The robot will be your friend, not your foe

It’s reductive to assume that an entire job title is tantamount to a single, teachable task. In all likelihood, new automation introduced to the workforce will either complement human work or will cause certain jobs to be redefined without being lost.

A 2015 case study by James Bessen surveyed the effects of automation through the introduction of ATMs in the ’70s.

After the ATM’s proliferation, banks started employing fewer tellers per branch – but there were actually, across the board, more tellers employed than ever before.

ATMs led to greater productivity and saved labour, all of which saved money. The bank took these profits and decided to vie for a greater market share by opening more bank branches. This led to more jobs and, furthermore, resulted in bank tellers having more meaningful jobs.

The development of ATMs led banks to re-assess the role of the bank teller. Banks concluded that tellers could be instrumental in what is called ‘relationship banking’. Tellers, instead of mechanically doling out cash, engaged with customers and devoted more time to selling financial products, such as mortgages and pensions.

While Autor does warn that automation is not always guaranteed to produce such results, the robot in your life in the future will likely be one that frees up your time from stultifying and predictable tasks, which probably weren’t particularly enjoyable to begin with.

The automation paradox – when robots go rogue

On 31 May 2009, 228 people died when Air France flight 447 bound for Paris crashed into the Atlantic. It was one of the deadliest crashes in Air France’s history – and one that was, investigations concluded, both directly and indirectly caused by automation.

While autopilot is an extremely common feature on modern aircrafts, the aircraft these pilots were flying was exceptional. The AirBus A330-203 has an additional layer of automation, which means that not only does the plane keep itself steady, but it regulates the controls using a ‘fly-by-wire’ system.

When a pilot moves a stick forward to indicate that the plane needs to change angle, AirBus automatically adjusts the inclination so as to keep the correct amount of air flowing over the wings, which maintains the amount of lift needed to keep the plane in the air.

Forget yourself, forget your skills

The plane’s sensors froze over and, when this happened, the plane could no longer regulate itself, unbeknownst to its pilots. The plane quickly began to stall, meaning it no longer had enough lift, and began to descend.

Its pilots, stupefied, had no idea what was going on despite the word ‘stall’ being played loudly on repeat over the cockpit speakers. As journalist and former pilot William Langweische described it, “The pilots were hideously incompetent.”

The model of plane they were flying was designed by a man named Bernard Ziegler, who once boastfully proclaimed that he was going to build a plane that his ‘concierge’ could fly.

This proclamation denigrates the value of human input, and it was this denigration that lead to disaster. Of the 346 hours of flying the captain had logged over the six months before the crash, he had only spent four hours in control of the plane. The rest was left to automation, meaning that the pilots’ skill-set began to decay so severely that they couldn’t recognise an issue that would have been discussed in the first days of pilot training.

Overestimating the technology, underestimating ourselves

One of the biggest risks associated with new developments in automation is that human capabilities will be undervalued in conjunction with AI being overestimated. This was very grimly demonstrated last year with the first confirmed Tesla crash. The car’s occupant allowed autopilot to reign supreme only to be killed by its failure.

If history is any indication, society will not descend into ruin when new technology comes around. If anything, one of the most ruinous things we could do is be too quick to abdicate the responsibility, assuming that AI will be able to pick up the slack. Dreamy predictions of a work-free world can make it that people are too willing to prematurely trust the technology regardless of whether it is sufficient.

What do we most have to fear? Wealth inequality

The robots aren’t going to take your jobs. It’s possible, however, that the wealth inequality, which is taking hold amid a very sudden tech boom, will pose a significant risk to our economy. It is now possible for two people in a garage to put their heads together, design revolutionary software, and generate billions in revenue. The productivity-to-employee ratio is, in that case, massively skewed. Upward mobility is under threat in a way that it hasn’t been since the 19th century, according to French economist Thomas Piketty in his 2013 book Capital .

This is not the technology’s fault. What’s truly to blame is the lack of public policy in place to create enough wealth redistribution to prevent a gulf in riches opening up in the coming years.

Do not point the finger at the new invention, but point it instead at whomever is allowing the inventor to hoard the spoils.

Eva Short
By Eva Short

Eva Short was a journalist at Silicon Republic, specialising in the areas of tech, data privacy, business, cybersecurity, AI, automation and future of work, among others.

Loading now, one moment please! Loading