Managing Automation: Employment, inequality and ethics in the digital age

Managing Automation

Managing Automation

Yesterday (28th December 2017), the IPPR (a UK think-tank) published their report into the effects of automation on the economy and posited some suggestions as to initiatives that might be put forward to address the problems foreseen with managing automation.

While the report comes on the back of a wave of activity during 2017 on the subject “Will Robots Steal our Jobs”? It’s refreshing to read a piece that considers a range of factors such as capital v income inequality as well as the ethical regulation of technology and technology companies.

For those who don’t have time to read the full 56 pages, this article is aimed at the lay reader who is seeking a more measured view than The Sun’s “Jobs Terminatored” headline which followed the publication of the report. Where appropriate, the report’s findings have also been critiqued.



The report is split into five sections, initially reporting the findings of the analysis conducted by Matthew Lawrence, Carys Roberts and Loren King of the IPPR and then moving on to suggesting measures that can be taken to avoid some of the pitfalls they foresee by the current trajectory of this technology.

The tone of the report can be summarised as “automation will transform the workplace – we need to act now to ensure the impact is fairly distributed”. This is predicated on a lengthy analysis looking at the effect of automation on the workforce and on unemployment rates more generally.

The conclusion of the analysis is that mass unemployment is a highly unlikely scenario, although makes much of the 2017 McKinsey report that indicates as many as 60% of occupations have at least 30% of activities that are automatable.

The article also seems to support Artz’ 2016 findings that 10% of UK jobs are at a high risk of automation and that by the mid-2050s (which, to put into context will be when my son will be the age I am now currently) half of current work activities in the global economy will be automated – and that this is likely to be true also for the UK.

It’s hard to square these findings with the toned-down response that “jobs will be transformed, not eliminated”. Even taking the most optimistic view of the future that the IPPR hold, that only 10% of jobs will be eliminated through automation, then this would be a much greater number than the 200,000-strong UK mining industry that rose up against Margaret Thatcher in the early 1980s. Reference to this point would have been worth noting in the report, as the long term effects of the transition of these workers is still being felt today.


Hot-dog, anyone…?

The optimism the authors feel for the limited (read: 10% unemployment) effects of automation on the economy is that increased automation actually might increase employment in other industries as automation will lead to higher productivity, higher wages, and therefore higher demand for consumption.

This same analysis was neatly captured in an analogy painted by Paul Krugman back in 1997 which he imagines a hot-dog stand that has three types of workers: Those that prepare the sausage, those that bake the bun, and those that put the sausage in the bun.

Imagining that the company invests in automation to do one of these functions (say preparing the sausage), the result isn’t simply the unemployment of the sausage producers, nor is it the suppression of wages as the former sausage producers move into baking and bun filling.

Instead, the result is likely to be that the cost of hot dogs comes down, this is reflected in the price of hot dogs as rival stands also start to automate – and the lower cost of hot dogs creates a surge in demand. Therefore, the company might actually hire more people in aggregate than it did initially to fulfil the two tasks as yet unautomated.

While this might seem like a rosy rationale to explain why technology hasn’t seemed to affect employment numbers to any great degree over the past half-century, it falls short on two levels:

Firstly, on an economic level – the hot-dog analogy only maintains employment levels or increases the levels of employment when there is latent demand for hot-dogs in the economy.

Imagine the price of hot-dogs has been reduced so much that the cost of production is now trivial, and hot-dogs are no longer a treat but are in abundant supply.

Will demand for hot-dogs continue to rise inversely proportioned to their cost?


There comes a saturation point where hot-dogs reach a peak consumption, and the rise of employment in that industry will remain static.

This can be seen say in the fast-food industry (employment levels in this industry peaked with market saturation), as with the mobile-phone industry (again, employment levels in their production peak with market saturation).

Secondly, the hot-dog analogy fails in the face of technological reason. It assumes that as human labour retreats to ever-safer ground, the pace of automation will plateau or cease. This is based on a 20th Century view of technology automation, which is limited to viewing technological progress being focussed in ‘super-human’ abilities and human labour moving into those jobs requiring human level abilities.

Looking at the hot-dog van, it might well be the case that the preparation of the sausage and the baking of the bun might be tasks that a machine can do better than a human because an efficient process can enable the parallelisation of sausage processing in a way that a human simply couldn’t deliver.

The putting of the sausage in the bun however, might require a level of dexterity that robotics simply can’t deliver and therefore that’s the point where automation ceases.

What I’ve described here is a turn of the millennium view of the world.

The advances in Artificial Intelligence, Automation, and Robotics since then have been exciting (and troubling) because they are exactly in the field of human-abilities.

As I said in my recent TEDx talk, “Robots are being given dexterity. Things that you and I take for granted, like the ability to turn the handle of a door, open, and walk through it – will soon be something a machine can achieve”. 

Whereas peak employment in both the fast-food and mobile phone production industries is reached at the same time as peak consumption – employment rates will reduce as automation continued to bite.

So where does this leave employment?

Well, the only jobs where there will remain a need for human labour will be those jobs that specifically require human touch. The trouble is, there just aren’t that many – and likely not enough to satisfy our current society’s predication on full employment economics.

In the hot-dog van analogy, imagine there is a fourth job – the one of the hot-dog seller, whose job it is to smile at customers and hand over the hot-dog.

Will there be enough demand for hot-dogs for us all to move into the ‘lovely job’ aspect of hot-dog customer service? I doubt it. At least the positive side of this story is that the ‘lousy job’ aspects of hot-dog production will be consigned to history.


Distribution of Automation

The second section in the report deals with the distribution of automation across the economy.

The report focusses on the likelihood that those in low-income employment and therefore low-skill work are likely to be affected by automation more sharply than those in middle-income or high-income work.

The risk, of course, we want to guard against is that those who have limited resources, either capital, income, or skill should have their interests protected, but we also need to take a measured view across the entire economy.

The likelihood that professions that society has held sacred such as lawyers, accountants and doctors will be decimated by automation is a point too important to miss.

While referencing Susskind and Susskind’s work “The Future of Professions”, the report fails to address their central point; that the organisation of professional training and practice that we have today is a paradigm of the ‘print-based economy’, and simply is not fit for purpose for the ‘digital information economy’.

In my own work I’ve noted that it’s the low-skilled or low-income worker who is much less in denial about the effect of automation on their career than the doctor, lawyer, or management consultant.

It’s also the case that while those in low-skill or low-income jobs might need support to retrain, but likely will manage the transition successfully – those in professional careers who are affected by automation are at risk of being unemployable.

Seven years in Med-School and two decades in specialised practice doesn’t necessarily give you highly valuable transferable expertise once your field is being automated. Managing the transition for the professionals once they face the reality of their futures is going to be a conversation society needs to have. And soon.

While the report does note the fact that the UK is currently a leader in this space (London is a world-centre for AI technology, and Bristol is a wold-leader in Robotics), it does not foresee the effect of overseas capital and control on industry.

I’ve written on this point before, but to summarise – in automation there are winners and losers. One group of losers is clear. It’s the people who once did jobs which are now being automated.

The related group of winners is equally clear, it’s the former employers to those workers. What’s less obvious, and therefore often forgotten is that there is a second group of winners and losers – the effect of automation on which is much more drastic.

The winners are the companies that own the intellectual property of the automation technology which is being licenced to the companies that have replaced their workers with machines.

The losers are the nations who previously raised taxation from the income of the workers who are now unemployed, and which are likely not the same jurisdictions as those who can tax the capital owners of the technology.

The effect of this is huge. States need to support workers in transitioning to new jobs, while at the same time balancing the budget. It’s like supporting your kids’ education again, at the same time as taking a drop in salary.

At a national level. Much is made of London and the UK’s leadership in AI technology, but the fact is that while the talent might be located in the UK (and will continue paying taxes here) – the source of Capital is from overseas.

We’ve witnessed the difficulty regulators like TfL have had with foreign entrants like Uber in simple matters of following existing legislation. Imagine if TfL tried to windfall tax Uber in order to compensate the taxi drivers put out of work? I think we’ll see exactly this story play out in the coming years.


Accelerating Automation

I was delighted to see the IPPR advocate accelerating automation in order to reap the benefits of productivity improvement. However, the report doesn’t see the need for selective acceleration of automation in essential industries at the same time as throttling it for others.

The authors of the report appear to be of the view that the effect of automation will be linear and continuous. However, automation is exactly the Black Swan paradigm of discontinuous change that Nassim Nicholas Taleb warns us about.

Perhaps the last Black Swan in tech automation was the Spinning Jenny, and so the effects have been forgotten – but the reality is that automation is likely to hit the economy much as the waves hit the shoreline.

What we’ll likely see over the next century is as technology advances to automate a greater amount of human ability, wave after wave of jobs will be eliminated.

The effect will be culminative.

The task of the ‘Productivity UK’ body whose creation is suggested by the authors will be to balance the impact of automation on some industries at the same time as incentivising automation of others.

I suggested in my recent TEDx talk that the way to do this would be to incentivise the automation of labour in industries that are essential for human survival, such as food production and distribution.

Doing so would offset the effect of loss of income on those displaced. This subject and that of ‘Robot Taxation’ is something I intend to write about more in the New Year.

The second issue caused by a myopic focus on productivity, is that we lose sight of what is truly important. Rather than paraphrase Robert Kennedy further, here is his 50 year-old quote in full – especially relevant today:

“Too much and for too long, we seemed to have surrendered personal excellence and community values in the mere accumulation of material things. Our Gross National Product, now, is over $800 billion dollars a year, but that Gross National Product – if we judge the United States of America by that – that Gross National Product counts air pollution and cigarette advertising, and ambulances to clear our highways of carnage. It counts special locks for our doors and the jails for the people who break them. It counts the destruction of the redwood and the loss of our natural wonder in chaotic sprawl. It counts napalm and counts nuclear warheads and armored cars for the police to fight the riots in our cities. It counts Whitman’s rifle and Speck’s knife, and the television programs which glorify violence in order to sell toys to our children. Yet the gross national product does not allow for the health of our children, the quality of their education or the joy of their play. It does not include the beauty of our poetry or the strength of our marriages, the intelligence of our public debate or the integrity of our public officials. It measures neither our wit nor our courage, neither our wisdom nor our learning, neither our compassion nor our devotion to our country, it measures everything in short, except that which makes life worthwhile. And it can tell us everything about America except why we are proud that we are Americans.”


Ethical AI

The simplest section of the report is also the one which I find the least issue with.

The authors advocate the establishment of a national body to pro-actively manage the ethical, legal, and regulatory challenges posed by the technology. This will include:

–         Creating a predictable liability framework

–         Privacy regulation

–         Safety regulation

–         Human-in-the-loop regulation

–         AI rights (see my first blog article on Robot Rights)

–         Military drones

–         Scrutiny of Machine Learning

–         Dealing with the ‘Trolley’ – Problem (see my upcoming blog article in the New Year)

–         Superintelligence, and presumably – the Singularity

The authors suggested modelling such a body on the Human Fertilisation and Embryology Authority, and such a suggestion appears sensible; although given the specific impacts of technology on the economy, but also on social and political debate (such as the recent Cambridge Analytica controversy) – it is surprising the authors did not push further on this notion and suggest that international co-operation is necessary, perhaps through the United Nations to mediate and regulate humanity’s relationship with technology, and sanction those who run ahead without regard for the rest.


Capital Wealth

Returning to the earlier points made regarding the winners and losers of automation, the report concludes with suggestions of a number of measures which might help address capital inequality.

1.     Establish a Citizen’s Wealth Fund

2.     The Expansion of Employee Ownership Trusts

3.     The Introduction of Compulsory Profit Sharing

4.     Reducing Working Hours

While all well-intended and worthy of further debate, what the authors again fail to recognise is the cross-border impact of the technological change foreseen and that this needs to be reflected in initiatives designed to counteract its negative effects.

For example, mandatory profit sharing will only benefit those who remain in employment – and doesn’t do anything to counteract the disproportionate capital wealth creation overseas.

What is needed are solutions that recognise these complex international relationships that are common in the digital economy, and do more to ensure capital is not just fairly distributed within a nation, but between nations.

Unless this warning is heeded, the balance of power is likely to continue to shift at an increasing rate towards Silicon Valley and away from our national democracies, and along with it our income and capital creating abilities also.

7 emerging technologies eBook

Interested in technology and the effects of technology on society?

Sign up and get a free copy of my ebook here: