Chris Middleton explains why the dominant narrative about robots equalling mass unemployment is wrong. However, the nature of work and skills is certainly changing.
In 2018, the World Economic Forum (WEF) predicted that artificial intelligence (AI), robotics, automation, and other ‘Industry 4.0’ technologies will transform employment. Its report, the Future of Jobs 2018, forecast that 75 million human jobs would be lost to the technologies. These would not only be in production-line processes, but also in once-safe white-collar careers, such as banking, financial advice, auditing, accounting, legal services, and administration – roles that are increasingly rules-based and repetitive, and so vulnerable to automation.
On the face of it, the figures appeared to give credence to tabloid stories about job-killing machines and mass human unemployment. Speaking at a Japan/UK Robotics and AI Seminar in London in 2016, Dr Anders Sandberg of the Future of Humanity Institute at Oxford University had already reinforced those fears by saying, “If you can describe your job, then it can – and will – be automated.” He suggested that nearly 50 percent of all jobs could fall to the machines.
But this was only part of the story reflected in the WEF report. In total, 133 million new jobs would be generated over the next few years, it said: a net gain of 58 million to the world economy. In short, Industry 4.0 technologies will create nearly twice as many jobs as they destroy, partly through productivity gains and the influx of new companies and services (as happened with ecommerce and mobility), and partly through the nature of work itself changing.
Out would go many rote, low-value-adding processes, and in would come more strategic, creative, and data management/analysis focused roles. Whether that means machines complementing the work of humans or vice versa remains to be seen, but the narrative that ‘robots in’ equals ‘humans out’ in a zero-sum game has long been unconvincing – partly because there is little hard evidence to support it.
By far the most automated country in the world is South Korea, which has a robot density of 631 for every 10,000 human workers. That’s nearly nine times the UK ratio of robots to humans; Britain has one of the lowest robot densities in the developed world, at 22nd in the global automation league. So if it’s true that more robots equals more humans out of work, then you would expect South Korea to have soaring unemployment, but as of August 2019, unemployment in that country stood at 3.1 percent – significantly lower than the UK (3.9 percent), after a spike earlier in the year.
China is automating faster than any nation on Earth and yet also has an unemployment rate lower the UK’s, at 3.6 percent. The point is made even more strongly in the case of Japan, which hosts 23 percent of all the world’s robots, including nearly 300,000 industrial machines, and yet has just 2.4 percent human unemployment.
Singapore (2.2 percent unemployment), Germany (3.1 percent), the US (3.7 percent), and Taiwan (3.7 percent) are also among the top 10 automators, and yet all have unemployment rates that are lower than the unproductive, largely un-automated UK. The exceptions in the world robotics top 10 are Italy (9.7 percent unemployment), Sweden (7.1 percent), Belgium (5.6 percent), and Denmark (4.8 percent), yet in each of these cases unemployment has generally been falling in the long tail of the global financial crisis.
Critics would rightly point out that such statistics can be misleading: they fail to capture the numbers of people in temporary employment, on zero hours contracts, in the gig economy, or on extremely low wages, and it’s conceivable that the influx of robotics and automation could create a new layer of menial, largely invisible workers. Youth unemployment is also high in several of those nations.
However, the point stands that there is no simple causal link between automation and mass unemployment in these countries – at least, not yet. Far greater stresses on the world economy and jobs include the China/US trade war, political instability in Europe, soaring debt, and the ongoing prevalence of high-risk investments in the financial markets.
But what about jobs in manufacturing specifically? According to recent PwC analysis of Bureau of Labor Statistics data, the most robot-intensive manufacturing sectors in the US, such as automotive, electronics, and metals, employ 20 percent more mechanical and industrial engineers than other manufacturing sectors – and pay them higher wages. In other words, robotics appears to be generating new skilled roles in factories, even as it replaces lower-skilled functions.
Meanwhile, 2017 figures from the International Federation of Robotics (IFR) revealed that US manufacturers purchased 60,000 industrial robots in 2015-16 – as did China – but 250,000 human jobs have been created in or brought back to the US manufacturing and supply chain sectors since then.
Amazon – one of the most highly automated and robot-assisted companies on Earth – has created tens of thousands of new human jobs in recent years. Often criticised for its treatment of fulfilment centre workers and aggressive business practices, the company has nonetheless added thousands of skilled employees, including 1,000 in the UK alone last year, bringing its total UK workforce to over 27,000.
As for the UK itself, the country faces flatlining productivity, low automation levels, and political instability over its fracturing relationship with Europe. While inward investment may be holding up for now, that may be asset-stripping and there is mounting evidence of companies pulling jobs, money, and/or assets out of the UK and relocating them in Europe.
Last updated in July 2019, Bloomberg’s Brexit impact tracker still makes sobering reading, and there have been further reports of closures, warnings, or relocations since the Summer. That has little to do with automation.
But Brexit aside, what is the ‘big picture’ future for the UK technology and jobs markets? A recent government-commissioned report, Automation and the Future of Work, observed: “The problem for the UK labour market and our economy is not that we have too many robots in the workplace, but that we have too few. In 2015 the UK had just 10 robots for every million hours worked, compared with 167 in Japan. By 2017, we represented just 0.6 percent of industrial robotics shipments.”
In order to help address flatlining productivity since the 2008-09 recession, the September 2019 report says that UK organisations should invest in robotics and AI and encourage SME adoption of the technologies, back the Industrial Strategy with greater investment in automation, and create a new Sector Deal for robotics.
However, it acknowledges that rising unemployment is a risk from such policies – one that should be counterbalanced by fostering a smarter, more supportive environment for skills, training, and education, both nationally and in enterprises. Failure to create such a culture risks deep knowledge haemorrhaging out of our organisations.
Meanwhile, a separate report repeats the still-dominant narrative that ‘new technology in’ equals ‘humans out’. It forecasts that up to one-third of UK jobs will be automated, or are likely to change as a result of AI alone. The document, Harnessing the Power of AI: The Demand for Future Skills, has been published by global recruitment firm Robert Walters and market analysts Vacancy Soft.
However, the report provides evidence that 2018 WEF forecasts about the changing nature of work were correct, as data volumes boom worldwide. Ollie Sexton, Principal at Robert Walters said, “As businesses become ever more reliant on AI, there is an increasing amount of pressure on the processes of data capture and integration. As a result, we have seen an unprecedented number of roles being created with data skill-set at their core.
“Our job force cannot afford to not get to grips with data and digitalisation. Since 2015 the volume of data created worldwide has more than doubled – increasing on average by 28 percent year on year. So now is the perfect time to start honing UK talent for the next generation of AI-influenced jobs. If you look at the statistics in this report we can see that demand is already rife; what we are at risk of is a shortage of talent and skills.”
Talk of a UK skills gap has been around for the past 25 years; yet according to the report, professionals dedicated to data management are now the fastest-growing area of recruitment. Within large or global enterprises, volumes have increased tenfold in three years.
Overall data roles have increased by 80 percent since 2015, with data scientist vacancies growing by 110 percent year on year and data engineers by 86 percent. Meanwhile, an explosion of cybercrime has forced services firms to hire information security professionals “aggressively”, says the report.
According to Robert Walters, the top industries Investing in AI are: (in descending order) Agriculture; Business Support; Customer Experience; Energy; Healthcare; Intellectual Property; IT Service Management; Manufacturing; Technical Support; Retail; and Software Development.
Tom Chambers, Manager of Advanced Analytics and Engineering at Robert Walters, said: “The uptake of AI across multiple industries is bringing about rapid change. Particularly, we are seeing Retail, Professional Services, and Technology industries strive to develop digital products and services that are digitally engaging, secure, and instantaneous for the customer – leading to huge waves of recruitment of professionals who are skilled in implementing, monitoring, and gaining the desired output from facial recognition, check-out free retail, and computer vision, among other automation technologies.
“Similarly, experimental AI is making huge breakthroughs in the healthcare industry, with the power to replace the need for human, expert diagnoses. What we are seeing from those businesses that are prepared to invest heavily in AI and data analytics, is they are already outperforming their competitors – and so demand for talent in this area shows no signs of wavering.”
All of that may be true. However, organisations should be wary of rushing in to implementing these technologies tactically and on a wave of hype, rather than strategically and in line with business goals. Claim after claim is made for AI’s abilities in particular, but the belief that it is inherently superior to human intelligence is a dangerous one.
Take healthcare, for example. There is no doubt that AI’s pattern recognition abilities may help to transform disease diagnosis and personal healthcare management in the near future. However, medical journal The Lancet recently published a critical, detailed, and systematic review of deep learning and AI in the field of medical imaging – excerpted and discussed in this Twitter thread by author Eric Topol.
One Twitter user responded, “This should really put to bed facile statements about AIs ‘beating’ clinicians at their jobs, even in the more advanced areas of medicine.” That user was none other than Michael Macdonnell, Director of Google Health, and formerly of DeepMind.
Strategy, deep reading of research, measurable results from technology programmes, and investment in skills: that’s the sensible route ahead for business leaders – alongside creating agility and the ability to capitalise on new opportunities quickly.
Noise, techno evangelism, and tactical responses are counter-productive and may simply lead to disappointment. The message is: Don’t retire your human intelligence for the foreseeable future, and resist losing experience and expertise from the organisation. Use technologies such as AI and automation to complement human skills and free up your employees to do what they do best.