Robo-pocalypse not-quite-Now

Chris Middleton Profile picture for user cmiddleton September 21, 2017
The jobs robo-pocalypse is postponed, says YouGov. 
But is the UK’s moderate approach to robotics the right idea, or economic suicide?

Robot and human connecting through electricity bolts © Pixelbliss -
Another day, another apocalyptic story  about robots stealing human jobs. This time, the headlines shout that over four million UK private sector jobs will be seized by machines, AKA 15% of the total workforce.

Unusually, the story is sourced from YouGov statistics, which are quoted in a new report, The Age of Automation, published by the RSA. However, closer reading of the 83-page document reveals the mainstream media headlines to be sensationalist and inaccurate.

That’s no surprise, but it does little to help organizations locate the signal in the constant media noise about a machine apocalypse.

What the statistics actually say is that a clear majority – 60% of the UK business and IT leaders questioned by YouGov – believe that the impact of AI and automation will be minimal to low. Twenty-two percent see “zero potential” for automation in their businesses, while 38% say that between one and 15% of jobs may be automated. That’s a completely different story.

But what of the other 40% of decision-makers? Twenty-seven percent of respondents believe that 16-30% of jobs may be automated over the next decade, while just 13% believe that the technologies’ impact could be higher, with over 30% of jobs being taken by the machines.

The YouGov findings stand in marked contrast to other UK studies, which have suggested that between 35 and 47% of jobs will be automated. The higher of those two estimates was made by Dr Anders Sandberg of Oxford University’s Future of Humanity Institute. Speaking at the Japan/UK Robotics and AI Seminar in February 2016, Sandberg said:

If you can describe your job, then it can – and will – be automated.

What sets the RSA report apart from the usual tabloid fodder is its suggestion that this is not the case:

Our first conclusion is that AI and robotics are more likely to alter jobs than to eliminate them. Despite impressive advances in machine capability, many tasks remain outside of their scope, particularly those demanding manual dexterity and deeper forms of creativity and communication.

Moreover, automation tends to be task-based rather than job-based, allowing workers to pivot into new roles should machines encroach on their turf. No single device can wholly substitute for retail assistants, care workers, hotel receptionists, or building labourers.

That’s true, and as diginomica's recent report on automation in the US manufacturing sector revealed, robots are creating new skilled opportunities, as well as stripping away low-skilled, routine tasks. This makes skills and rising economic disparity the key societal issues.

But the RSA ignores two things in reaching its conclusion: first, robots are just as likely to be software as they are hardware – such as the chatbots that could replace up to one million UK call center workers; and second, softbots tend to remove the first rung of the ladder towards professional careers.

Even so, uptake may be slower than the doom merchants predict. The report goes on to quote a Cisco and Capita survey of IT leaders, which found that while 50%  view AI as relevant to their organization, just eight per cent are currently putting it to use. In the case of robotics, the figures are 39% and 10%, respectively.

Yet despite the hype cycle we find ourselves in, these technologies are on an upward swing and they create strategic challenges for every organization, even if their uptake may not be as rapid as some predict.

So what are the barriers to faster adoption? The RSA identifies four: cost and business models; consumer preferences; regulatory concerns; and organizational integration. For example, the report discusses the healthcare market and other sectors in which specialist robots are expensive, rudimentary, and on a fast track to obsolescence – while incurring day-to-day maintenance fees.

This suggests that robots are like cars: depreciating assets, unlike humans who can constantly acquire new skills or set out to change their lives and employment opportunities – assuming that such opportunities still exist where they live.

But metaphors are dangerous, as Milan Kundera once observed. It’s wise to remember that the car itself will change over the next few years. The future is an app-summoned, rentable service provider that can update itself overnight with data about how every other car in the world has been used. This ability to gather and crunch vast amounts of data is what set the robots apart from human beings. Individual components might age, but their intelligence can only increase.

Despite this, the report pursues the ‘depreciating asset’ angle:

Why invest in a RIBA robot if there are rumours of a more sophisticated caring robot just around the corner? And why plough money into a fraud detection algorithm with 95% accuracy if there are expectations one will soon emerge with 99% accuracy?

Fair point, but that can be said of any technology, from cars, laptops and smartphones to enterprise applications, and it doesn’t prevent organizations from buying them. The critical question is why they are buying them: strategically and sustainably to satisfy a real business or societal need? Or tactically in an unsustainable arms race towards lower and lower costs? That’s the area that high-level studies need to focus on.

Of course, the risks the RSA identifies can also be mitigated in the cloud with software as a service (SaaS) agreements, in the case of AI or software that automates back- and front-office tasks. This on-demand, plug-and-play model is emerging in robotics, too, with Saviotic’s hotel concierge robots and Starship Technologies’ delivery droids now available on rental, rather than purchase, schemes.

The phone droid

That said, the thrust of industrial robotics, for example, is towards smaller, faster, customisable platforms that are driven by sector-specific apps. This obviates the need to manufacture the kind of cumbersome, depreciating, single-task machines that the RSA focuses on. In other words, think of the next generation of robots as being like scaled-up smartphones, which can carry out a range of complex manual tasks. The apps are the key; the rest is just speed, processing power, and dexterity.

Either way, organizations must reflect on their wider business strategy before buying in to these technologies, says the report. Decision-makers should weigh up the cost of a new technology versus the savings that could be made via traditional staffing or efficiency improvements:

For organizations employing well-paid and highly skilled staff, there may be an obvious case for buying in machine alternatives (one reason why the financial industry is bracing itself for significant disruption). However, for organizations operating in low-skilled and low-paid sectors, including care homes, restaurants, bars, and some factories, it will continue to be cheaper to employ people.

Again, this may be true – if one ignores the implication that skilled or highly paid humans are somehow a problem – but the RSA is well behind the curve. The financial services industry isn’t ‘bracing itself’ for disruption, it’s already in the vanguard of the technologies’ adoption worldwide. The banking sector has long been little more than an automated compliance program.

Ultimately, the RSA’s message is – correctly – that no technology is predestined to impact on society in the way that the doom-mongers predict:

As a society we have a choice in how to apply AI and robotics and manage their effects. […] These choices, however, are only relevant so long as the technology is being deployed. Indeed, just because a machine can do something, does not mean that it will be bought, integrated, and licensed to do so.

That’s true, and we must do our utmost to steer these technologies towards Sustainable Development Goals  and societal benefits for all.

My take

One finding stands out from the RSA report, and it should sound as a note of caution about all of the others: the UK must do better in this sector, or lose its chance of being a major player in the biggest new market this century.

The report says:

Herein lies the rub for the UK: while as a society we have been quick to lament the rise of AI and robotics, as an economy we have been slow to adopt these technologies.

The RSA reveals the extent of the problem in one sector alone. Far from capitalizing on industrial robots’ ability to make manufacturing smarter and more efficient, robot sales to the UK actually decreased in 2014-15, with the UK purchasing far fewer than the US, Germany, France, Spain, and Italy.

The report adds:

In 2015, the UK had just 10 robot units for every million hours worked, compared with 131 in the US, 167 in Japan, and 133 in Germany. While this may reflect our different sectoral make-up, UK businesses and public services as a whole suffer from stubbornly low rates of investment.

ONS data shows spending on gross fixed capital formation – a measure of investment that includes plant and machinery, software and new dwellings – has barely grown in real terms over the last decade. Going further back, data from the World Bank shows the proportion of UK GDP accounted for by gross fixed capital formation has fallen by seven percentage points since 1990.

These are alarming findings when set alongside the Global Innovation Index’s recent criticism of poor UK investment levels in education and the national infrastructure.

This leaves the UK in a fascinating quandary in the run-up to Brexit: is Britain’s moderate, ‘steady as she goes’ approach the ideal model in a world that is automating too fast? Or is the country on a fast track to economic irrelevance, investing too little, too slowly, and too conservatively in a transformative technology, and nowhere near enough in educating its people?

One thing is certain. Rewind 25 years and compare today’s YouGov predictions of a minimal impact of AI/robotics with many business leaders’ dismissal of the internet, ecommerce, or mobile phones. I remember interviewing record label CEOs in the mid-90s as they rejected the possibility of widespread disruption in their industry. I recall similar conversations across every sector: retail, insurance, leisure, transport, tourism...

Now imagine what the UK could have achieved had it properly invested in digital transformation on day one, or built Tech City in 1995, or backed Tim Berners-Lee with a billion pounds instead of a knighthood, or rolled out a nationwide fibre network when it had the chance and not left it up to BT's largesse to decide what's good for us all.

But then again, not even the US got everything right. Take this article from Newsweek in 1995:

The truth is no online database will replace your daily newspaper, no CD-ROM can take the place of a competent teacher, and no computer network will change the way government works.

That’s hindsight for you.

A grey colored placeholder image