Global Study: Robotics, artificial intelligence and work: hope for the best but prepare for the worst
- 04 Oct 2017
- James Davies
Firm: Lewis Silkin
Barely a week goes by without a new prediction about the impending impact of robotics and artificial intelligence (AI) on today’s jobs. There is, however, a striking lack of consensus among commentators about what the future holds.1
These are not new phenomena. Robots2 date back many centuries.3 AI4 is much newer, although is often said to have been born more than 60 years ago at the famous Dartmouth Workshop.5
The rapid development of technology is one of the three great drivers of workplace change in the early part of the 21st century. Crystal ball gazing is notoriously unreliable, but we can be reasonably confident that this pace of change will only increase. The other two main drivers, demographic change (particularly the rapidly ageing population) and globalisation, are also likely to continue and even accelerate. While the advent of the Trump presidency in the US and Brexit in the UK have suggested a backlash against globalisation in some quarters, these are likely to be mere hiccups in the long-term trend towards a more interconnected world.
Optimists versus pessimists
Commentators on the impact of robotics and AI on work largely break down into two schools: those who predict major disruption to jobs, and on the other hand, those who predict that the doomsayers are wrong and the impact on jobs will be positive.
The optimists argue that the world of work is merely undergoing a further phase in a process of constant evolution. They point to doom-laden predictions in other eras which proved false. Even as long ago as ancient Greece, Aristotle was arguing that if machines could become sufficiently advanced there would be no further need for human labour.6
During the First Industrial Revolution,7 agricultural jobs vanished but were replaced by manufacturing jobs. When manufacturing declined in “richer” countries in the second half of the 20th century, service industry jobs replaced them. Robotics and AI are heralded by some as being part of the Fourth Industrial Revolution,8 although the Third9 and Fourth Revolutions are arguably part of the same period of change in technology.
Optimists recognise that many current jobs which employ significant numbers will reduce significantly, but contend that they will be replaced by other jobs. They argue that technological advances over the centuries have always resulted in a net increase in jobs, resulting from the increased productivity that accompanies such innovations.10 Some optimists go further and argue that, at least in the UK, the economy is in desperate need ofmore robots, pointing to the high levels of employment coupled with poor levels of productivity to support this.11
New jobs could come from increased demand for those that already exist or the creation of entirely new occupations. We can predict that new jobs will be created in areas such as virtual reality, while others are no doubt beyond our wildest dreams. During my childhood in the 1960s and 1970s, the internet would have seemed as fanciful to me as Spock being beamed up by Scotty and transported to another planet in Star Trek. Job titles such as social media manager, app developer or search engine optimisation consultant would have been utterly incomprehensible.
Optimists also say that improved productivity from use of technology can result in increased leisure time. Such projections go back many decades, with John Maynard Keynes having famously predicted in 1930 that we would only be working 15 hours per week by the time his grandchildren were part of the labour force.12 I am old enough to remember predictions that the technological advances of the 1980s would lead to increased leisure time for all. In practice, for many, technology has instead resulted in being “on call” 24/7, never leaving work behind and experiencing increased stress.
Calls for reduced working hours nonetheless continue to be made. Google’s co-founder Larry Page, for example, has advocated a four-day working week.13
Impact on jobs
Optimists and pessimists generally agree that the skills sought in the future will be different from those valued today. Jobs predicted to be most at risk are those that require routine manual effort or the analysis of data. Frey and Osborne14 produced a report in 2013 estimating the probability of computerisation of 702 occupations. These ranged from jobs such as telemarketers and library technicians, with a 99% chance of computerisation, to dentists, social workers and primary school teachers with less than a 0.5% chance of being replaced.
Other commentators go further, such as The Guardian journalist Dan Tynan who recently set out how actors, teachers, lawyers and even journalists are at threat from AI.15
Caring jobs are among the jobs least at risk, according to Frey and Osborne. It is easy to see how the demographic changes resulting in an ageing population will lead to increased demand for nursing and other caring roles. Deloitte’s 2015 report16 found that between 1992 and 2014 there had been a profound shift towards the provision of care and education services,17 a trend that might also be partially explained by changes in family life.18
There is, however, no consensus that caring roles will continue to increase. One recent prediction went so far as to suggest that robotic nurses could be the future of healthcare,19 a development that is already happening in Japan.20
It has been suggested that Frey and Osborne may have overstated the impact of technology on jobs.21 Later reports have highlighted that relatively few jobs may disappear entirely, although almost none will be wholly unaffected. The activities which go to make up a job may become replaceable, even if the entirety of the work done by the human in that job might not be susceptible to automation.22
Arguments of pessimists
The pessimists argue that this time things will be different. They point to the unprecedented speed of change, allowing less scope for the evolution of new jobs to replace those which are disappearing. Among those warning of the need to prepare for a world of less human work are Tesla founder Elon Musk23 and British scientist Professor Stephen Hawking.24
Online shopping has hit the number of people working in stores. This has to some extent been mitigated by an increase in driving jobs as the roads become congested with delivery vans, yet businesses are already experimenting with drones and driverless vehicles.
Goldman Sachs recently predicted that job losses in driving jobs across the US will amount to 300,000 per year.25 What will replace them?
Considering how very few jobs are likely to remain unaffected by technological changes, it seems unduly complacent to dismiss concerns that mass employment is at grave risk on the basis that “it didn’t happen last time”. The optimists may yet be proved right, but it would appear negligent not to plan on the basis things might turn out differently this time around.
Will it be different this time?
The global economic trends accompanying today’s technological advances give reason to fear that their consequences for jobs will not be positive.
A consistent feature of past revolutions has been increased personal wealth resulting from productivity gains, with fresh jobs developing to exploit these newfound spending powers. The study by Deloitte26 mentioned above highlighted the growth of bartenders and hairdressers from the 1950s to exploit the disposable income created by greater efficiencies and improved productivity.
In contrast, the current changes may be arriving at a time when economic growth is not being matched by increased prosperity27. Thomas Piketty’s seminal work28 has warned of the dangers arising from a low-growth economy in which returns from capital outstrip those from human labour. He is one of many highlighting increased inequality in societies across the globe and the threat this poses.29
For the optimists, technological revolution is sorely needed to stimulate growth. For the pessimists, technological developments are not actually having that effect, and without growth, new jobs will not replace those being made redundant by machines.
What should be done?
It is impossible to predict accurately how the nature of work will change in the years ahead. But policymakers should look at steps that might need to be taken if the less positive future vision is correct and we are required to adapt to a world with fewer jobs. There is an urgent imperative to educate and train people to meet the challenges of the future.
Fewer human jobs will revive interest in social safety nets and job creation programmes. Governments, politicians, lawyers, academics and businesspeople have already begun floating more radical ideas to respond to the potentially dramatic reduction in demand for human labour including, for example, a universal income30 and even human job quotas.31
These developments suggest that a root and branch reform of tax policy is needed. Taxing human jobs through employer social security contributions (13.8% of the wage bill in the UK) is arguably becoming increasingly unsustainable. As humans compete with computers for work, our current tax system means that humans do so with one hand effectively tied behind their backs.
Employer national insurance contributions do, however, generate over £70 billion per year in the UK, around 10% of the government’s entire tax receipts. The money will have to come from somewhere. There is already discussion of a wealth tax which could not only replace lost revenue from work-related taxes but also address the increasing inequalities mentioned above.32 Others have proposed a tax on robots.33
A failure to prepare for these changes could result in social unrest.34 Increased inequality and stagnating wages were arguably at the root of the Brexit vote in the UK and Donald Trump’s election as US president. In both cases, globalisation was apparently rejected by a large proportion of the population. One does not need to be clairvoyant to foresee further “Luddite”35 backlashes from those left behind by the technological progress.
The impact of robotics and AI is unlikely to be geographically neutral. Another predictable change is the transfer of “work”, even if undertaken by robots rather than humans, from locations with low labour costs to those with strong education systems. Clearly, the economies of cheap labour become less relevant where work is roboticised. Countries such as Singapore, South Korea, Hong Kong and Taiwan are likely beneficiaries of this shift while the losers may be nations in the developing world, resulting in increased geopolitical tensions.
Frey and Osborne have recently built upon their predictions about the effect of technology on jobs in the US by looking at the impact globally.36 They suggest that while 47% of jobs are vulnerable in the US, the numbers rise to 69% in India and 77% in China.
The role of law
A recent contribution to the debate was a report published in April this year by the International Bar Association Global Employment Institute (IBA),37 which sets out a thorough account of the future challenges of this Fourth Industrial Revolution. The authors, led by renowned German employment lawyer Gerlind Wisskirchen, remind us that while experts disagree about how long we will wait for these changes, they all agree on their inevitability.
The report emphasises the challenges for employers and lawmakers in adapting to the increased flexibility of working arrangements these technological developments herald, whether in relation to working hours, location of work or remuneration systems. The authors also highlight the new forms of employment that are springing up, resulting in an erosion of the dividing line between employment and self-employment,38 a topical issue with the growth of digitally-enabled independent work, often referred to as the “gig economy”.39
In the UK, the government commissioned Matthew Taylor to review how employment law was meeting the needs of modern working practices. His report40 highlighted the need for the law to keep up with workplace changes and address the needs of so-called gig economy workers. Employment courts around the world have been grappling with the employment status of Deliveroo riders, Uber drivers41 and the like, invariably applying legal rules that were designed for the workers of a bygone era. Minimum wage and working time laws established for more traditional types of work do not translate easily to individuals performing their work via online platforms, where the distinction between working and non-working time is more blurred.
A further area to which the IBA report devotes attention, which will pose increasing issues for legislators and businesses, is the growing importance of big data and its increased role in the workplace.
Employment law as we know it today developed in the UK in the 1960s with individual rights such as unfair dismissal protection and redundancy pay. These reforms coincided with the decline of manufacturing and a reduction in 9-to-5, full-time jobs.42 Nowadays, these rights based on protecting job security are coming under scrutiny as they become less appropriate for more flexible and varied working relationships enabled by technology.
The future is more likely to focus more on the human rights of workers, such as privacy, family life and the right not to be discriminated against. As the law develops in this direction, building on more recent legislation, the distinction between employment and self-employment will become less significant and necessary.
Regardless of whether sufficient new jobs are created to replace the redundant ones, the authors of the IBA report add to the call for employers, educators, legislators and policy-makers to prepare people for significant changes ahead. They persuasively conclude as follows:
It would be desirable for the future laws, which will hopefully be secured at the inter-national level by uniform standards, to be geared to the technological developments and the increased need for flexibility.
For employment lawyers like myself, it seems there will be work for some time. The Frey and Osborne study mentioned above,43 predicts that there is a 3% chance of a solicitor being replaced by a robot or computer compared to a 95% chance of an accountant being replaced.
As we do not seem close to any consensus as to what the future holds, it is also safe to say there will be plenty of work for futurists and commentators on the impact of robotics and AI on work. In the words of Benjamin Disraeli in 1833, “I am prepared for the worst but hope for the best”.
1. For examples of the ranges of predictions, see A Smith and J Anderson, AI, Robotics, and the Future of Jobs, 6 August 2014, www.pewinternet.org/2014/08/06/future-of-jobs.
2. “A machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer”: Oxford English Dictionary, robot, 21 September 2016, https://en.oxforddictionaries.com/definition/robot.
3. See F Latxague “A Brief History of Robots” Paris Innovation Review 18 July 2013.
4. “The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages”: Oxford English Dictionary, artificial intelligence, 18 September 2016, https://en.oxforddictionaries.com/definition/artificial_intelligence.
5. 1956 Dartmouth Summer Research Project on Artificial Intelligence at which many of the leading thinkers of the time participated.
6. R Campa “Technological growth and unemployment: a global scenario analysis” (2014) 24(1) Journal of Evolution and Technology 86.
7. 1760–1840, the development of mechanical manufacturing processes.
8. K Schwab “The Fourth Industrial Revolution” Foreign Affairs 12 December 2015. The Second Industrial Revolution is said to have taken place in the later 19th century and early 20th century with the advent of electricity, railways and mass production.
9. The Third Industrial Revolution is underway and takes account of the digitisation of manufacturing (see “A third industrial revolution” The Economist, 21 April 2012, www.economist.com/node/21552901).
10. For example, I Stewart, D De and A Cole Technology and People: The Great Job-Creating Machine (August 2015) www2.deloitte.com/uk/en/pages/finance/articles/technology-and-people.html.
11. B Dellot, Too Many Robots? The UK Doesn’t Have Enough, 25 July 2017, www.thersa.org/discover/publications-and-articles/rsa-blogs/2017/07/too-many-robots-the-uk-doesnt-have-enough; A Corlett Robot wars: automation and the labour market Resolution Foundation briefing (July 2016) www.resolutionfoundation.org/app/uploads/2016/07/Robot-wars.pdf.
12. JM Keynes “Economic Possibilities for Our Grandchildren” in Essays in Persuasion Macmillan, London 1931.
13. Interview with Vinod Khosla, Fireside chat with Google co-founders, Larry Page and Sergey Brin, 3 July 2014, www.khoslaventures.com/fireside-chat-with-google-co-founders-larry-page-and-sergey-brin.
14. CB Frey and MA Osborne The Future of Employment: How Susceptible Are Jobs to Computerisation Working Paper (September 2013) www.oxfordmartin.ox.ac.uk/downloads/academic/future-of-employment.pdf.
15. D Tynan “Actors, teachers, therapists — think your job is safe from artificial intelligence? Think again” The Guardian 9 February 2017.
16. Above n 10.
17. For example, 909% increase in nursing auxiliaries and assistants; 580% increase in teaching and educational support assistants; 183% increase in welfare, housing, youth and community workers; 168% increase in care workers and home carers.
18. See Organisation for Economic Co-operation and Development Doing Better for Families (2011) www.oecd.org/els/soc/doingbetterforfamilies.htm.
19. Research of Dr De Momi: M Banks, Robotic nurses could be the future of healthcare, 13 February 2017.
20. “Japan’s nursing facilities using humanoid robots, IT to improve lives, safety of elderly” The Japan Times 18 May 2017.
21. In R Berriman and J Hawksworth “Will Robots Steal Our Jobs? The Potential Impact of Automation on the UK and other major economies”, UK Economic Outlook, March 2017, the predictions of Frey and Osborne are compared with those of Arntz, Gregory and Zierahn (OECD, 2016) as well as the latter predictions re-modelled by PricewaterhouseCoopers. The AGZ and PwC predictions show less disruption than Frey and Osborne.
22. J Manyika, M Chui, M Miremadi, J Bughin, K George, P Willmott and M Dewhurst A Future that Works: Automation, Employment, and Productivity (January 2017). See also research of Oxford Institute of Humanity: C Edmond, This is when a robot is going to take your job, according to Oxford University, 26 July 2017.
23. Interview with MIT students: S Gibbs “Elon Musk: artificial intelligence is our biggest existential threat” The Guardian 27 October 2017 www.bbc.com/news/technology-30290540.
24. “Full artificial intelligence could spell the end of the human race” — interview with the BBC: R Cellan-Jones Stephen Hawking warns artificial intelligence could end mankind, 2 December 2014.
25. Goldman Sachs Economics Research (as reported by CNBC: A Balakrishnan, Self-driving cars could cost America’s professional drivers up to 25,000 jobs a month, Goldman Sachs says, 22 May 2017).
26. Above n 10.
27. See Organisation for Economic Co-operation and Development data for the UK referred to in V Romei “How wages fell in the UK while the economy grew” The Financial Times 2 March 2017; and for the USA, D Rotman “How Technology is Destroying Jobs” MIT Technology Review 12 June 2013.
28. Thomas Piketty Capital in the Twenty-First Century Belknap Press, Massachusetts 2013.
29. See also E Brynjolfsson and A McAfee Race Against the Machine Digital Frontier Press, 2011.
30. Finland is experimenting with 2000 young unemployed being given a basic income to study its effect on behaviours. Elon Musk is an advocate.
31. GWisskirchen, BT Biacabe, U Bormann, A Muntz, G Niehaus, GJ Soler and B von Brauchitsch Artificial Intelligence and Robotics and Their Impact on the Workplace, (April 2017).
32. For example, the Shadow Chancellor John McDonnell before the last UK general election.
33. For example, the Socialist presidential candidate Benoît Hamon in recent French elections.
34. Above n 1, at 53.
35. The Luddites were bands of English workers in the 19th century who destroyed machinery, especially in cotton and woollen mills, which they believed was threatening their jobs.
36. CB Frey and M Osborne Technology at Work v2.0: The Future Is Not What It Used To Be (January 2016).
37. Above n 31.
38. In the UK, the number of self-employed increased from 3.8 million in 2008 to 4.6 million in 2015 and the number of part-time self-employed increased by 88% between 2001 and 2015, Office of National Statistics, July 2016.
39. “A way of working that is based on people having temporary jobs or doing separate pieces of work, each paid separately, rather than working for an employer”: Cambridge English Dictionary, gig economy, 25 May 2016, http://dictionary.cambridge.org/ dictionary/english/gig-economy.
40. M Taylor Good Work: the Taylor Review of Modern Working Practices, (July 2017).
41. Cases brought in the following jurisdictions among others: UK, California, Massachusetts, New York, Alaska, Brazil, France, Canada, South Africa, and Switzerland.
42. In the UK, the proportion of the workforce employed in manufacturing declined from over 35% in 1961 to under 10% in 2011, Office for National Statistics.
43. Above n 14.