In March 2017, a study published by consultancy firm PwC predicted that 30% of jobs in the UK were under threat because of breakthroughs in artificial intelligence (AI), with up to half the positions in some sectors, like wholesale and retail, for example, likely to disappear.
This is not the first such study which has painted a grim picture of the future of labour markets because of the impact of AI. A widely noted study by Carl Benedict Frey and Michael Osborne in 2013 found that 47% of all American jobs were at threat from potential automation, whilst a Citibank 2016 report, published in collaboration with Oxford University, put the figure at 77% in China, and 59% across the whole OECD (Organisation for Economic Co-operation and Development) region.
Even world-renowned physicist, Stephen Hawking has weighed in on the debate, claiming in an article in The Guardian that AI would lead to a decimation of middle class jobs, worsening economic inequality and greater political unrest.
So, are we all about to see robots taking over from us, millions made redundant as the robots take over in some kind of post- Huxleyian nightmare?
The likelihood is that the future will be more nuanced than that. Whilst undoubtedly some jobs will be lost to AI, others will be created. There are also some service positions that are hard to automate, such as education, health and social care, hospitality and sports, and which are all expected to see healthy growth in employment levels.
Most at risk are those involved in jobs involving manual and repetitive tasks, in sectors like manufacturing or distribution. Workers in transport and logistics, such as taxi and delivery drivers, and office support, for example, receptionists and security guards, are likely to be replaced by computer capital, whilst those employed in sales and services – counter and clerical staff, telemarketers and accountants – are also in danger.
It should briefly be noted that changes in job market fundamentals due to automation is nothing new. The industrial revolution of the 19th and early 20th centuries saw the destruction of many cottage industries and rural employment, as labour began to be organised around mass production lines in factories and industrial centres. Subsequent developments in assembly-line production and technological innovation then saw the elimination of many manual labour jobs, whilst the advent of computers and the internet again had a radical impact on the pattern of employment, accounting for the loss of millions of clerical and administrative roles.
Despite this there has been no jobs “Armageddon” across the world, although it is undoubtedly true that many traditional manufacturing heartlands, such as the Rust Belt in the US and the North-East of England, have become economically depressed as local jobs have disappeared. In fact, what happens is that labour markets reposition themselves in the wake of such radical shifts, with new jobs created, as a consequence, in different sectors.
It also should be emphasised that changes in job markets tend to happen slowly, and this is likely to be the case with AI as well. People in sectors of employment identified as “at risk” should not start looking over their shoulders just yet.
In fact, the increased implementation of AI is likely to create a whole new swathe of jobs, requiring completely different skills and knowledge than exist today. A recent Accenture study “How Companies are Reimaging Business Processes with IT” identified 3 categories of new AI-driven business and technology jobs, which they labelled “trainers”, “explainers” and “sustainers”.
Trainers will be required to teach AI systems how they should perform – for example, helping language translators and natural language processors to make fewer mistakes, or teaching AI algorithms how to mimic human behaviour. AI systems do not understand human emotions or expressions of thought like empathy or sarcasm – they have to be taught.
Explainers will be needed to bridge the gaps between systems and humans, particularly business leaders and managers. Many business executives are uneasy with the “black box” approach of machine learning algorithms, and need somebody to explain and interpret data for them. Explainers will need not only to understand how the technology works and how it produces results, they will also need to be able to translate that into normal “business speak”.
The final category of job identified, sustainers, will be responsible for ensuring that AI systems are operating as designed. Current studies indicate that, on average, less than a third of companies will AI systems believe in their fairness, whilst there are serious doubts about their safety as well. This is a clear threat to the long-term roll-out of AI technology. Sustainers will have a big role to play in building the necessary trust and reliance on such systems going forward.
Of course, such jobs will require staff with the necessary skills and talents to undertake them, which, in turn, may need a radical shift by schools and universities away from traditional curriculum to offer courses which teach these new disciplines.
Undoubtedly AI is going to have an impact on the job market, but it is unlikely to be the doomsday scenario some have predicted. Those jobs which require a high degree of manual, repetitive tasks are likely to be replaced by computers or robots but, in their place, there are likely to arise a whole new range of jobs which are more highly skilled, challenging and crucial to long-term economic prosperity. And there will continue to be jobs which computers just cannot do, or perform well.
AI will have a transformative effect on employment, but it need not be cataclysmic.