A 2015 BBC article titled ‘Will a robot take your job’ certainly suggests that a jobless future is closer that many think, claiming that 95 different occupations in the UK have an 80% or greater probability of being taken over by machines in the next 20 years. The interactive page and search feature is based in part on a research paper by Carl Benedikt Frey and Michael A Osborne which claims to offer updated analysis of the problem based on recent advances in machine learning and mobile robotics. Machine learning has already seen widespread implementation, from Netflix’s and Amazon’s algorithms to forming the backbone of the company Rebellion Research, ‘the machine learning asset manager’. Such technology continues to advance rapidly, with the Google Brain team recently developing an AI programme capable of writing machine learning code to a higher standard than humans. Asides from the inevitable disappointment faced by children currently aged 1 or 2, as they enter the labour force to find their dream job as a telephone salesperson or ‘routine inspector and tester’ taken by intelligent machines, there are real issues facing societies and economies relating to the potential for widespread automation in the not so distant future.
One proposed solution to mitigate the problems arising from high unemployment due to the automation of jobs is a universal basic income (UBI) paid by the government to all citizens of a country regardless of them being employed or unemployed. There are many suggested benefits to such a scheme from reducing the exploitation of workers on low wages, reducing poverty, increasing charitable and voluntary work and providing people with greater freedom through a reasonably substantial safety net. A trial of such a scheme on 250 people in the Netherlands this year will offer each person £830 per month as well as an additional £130 for carrying out voluntary community work. Although currently the discussed benefits of such welfare provisions are centered around improving standards of living, happiness and freedom by giving people the security to pursue non-work related activities that may still be beneficial to society; with the potential for the future automation of such a large proportion of jobs, UBI schemes may be discussed in terms of them being necessary ways to distribute wealth to those unable to find work. The financial burden of a UBI programme is substantial, with the cost of giving each US resident $1,000 per month estimated at $4 trillion a year, there are however plausible arguments for how to fund it. Ideas proposed in the Netherlands include increasing tax rates on company profits to 30% as well as increasing taxes on large sums of personal savings. The idea of funding a UBI through increased taxes on company profits may be even more feasible in the future, if advancements in robotics, software and AI lead to a shift to capital intensive production and lower the costs of production of firms. In this case, the companies would effectively be taxed on the benefit they received in terms of increased profits from firing human workers and replacing them with machines, with this tax revenue being spent on providing income to the workers made redundant. Furthermore, the whole cost of UBI isn’t in addition to current government spending as it would replace various welfare programmes currently in place, in doing so reducing bureaucracy and the costs associated with it.
An alternative proposed solution to providing incomes for the significant unemployed section of the population involves an even greater degree of government intervention – this time in labour markets. This idea involves governments forcing companies to take on additional workers they don’t require at set wage rates, with the number of extra workers they must employ being linked to the company’s savings or profits. Even if these people weren’t making particularly meaningful contributions to their employers, they would still gain the sense of purpose and identity associated with employment, as well as a decent wage set by the government. This would again internalise the externalities associated with high unemployment by forcing companies whose profits increase as a result of automation to pay wages to workers they don’t need, therefore increasing the private cost of firing someone. However rather than grudgingly accepting reduced profits and higher costs, companies are more likely to offer training for these new employees and try to find roles for them to maximise their productivity and minimise their burden on the firm. This then potentially solves the issues arising from machines taking human jobs, by forcing firms to create new roles for humans to fill, potentially ones that robots can’t yet do to the same standard.
These ideas suggest that the future may not be so bleak for those 11 million British workers facing redundancy, with government intervention potentially reducing the costs to humans of producer’s drive for greater productivity and profits through capital intensive production. Some, however, have come to question the premise of the debate entirely. A research paper commissioned by the OECD to examine ‘the future of work’ found that job automatability may be overestimated by some studies as they assume that whole professions or large swathes of professions are taken over by programs or machines rather than just some roles within fields. This OECD research claims to account for the heterogeneity of roles within the workplace, and estimates that on average 9% of jobs in OECD countries are highly threatened by automation. The paper by Melanie Arntz and Terry Gregory also finds that there are likely to be variations within OECD countries as to how vulnerable populations are to unemployment linked to widespread automation. This figure is estimated to be just 6% in Korea compared to 12% in Austria. For the UK however the percentage of people at high risk of losing their jobs to automation is below 11%, significantly lower than the 35% previously suggested. Furthermore, technological change has the potential to generate jobs in new industries, and increased demand for technological products such as robots and software will increase jobs in the manufacturing, marketing and distribution of those products. The report concluded that any of the suggested changes in the labour market would be gradual due to legal, societal and economic hurdles.
In conclusion, then, there is a logical threat posed to human employees, especially those in low skilled jobs (and therefore in developing countries) by rapid advancements in technology that threaten to surpass human abilities in the workplace. Despite this, apparently worrying figures about the “automatability” of jobs may be misleading and miscalculated, and not paint the full picture of potential changes in the labour market. Although there are many forces resisting technological change and its displacement of human workers, even in a future with a greatly reduced supply of jobs for people, there are plausible solutions to support these people. Through government led intervention in terms of universal benefit payments or through holding firms to account for job losses, the negative impacts of technology on human wellbeing may be minimised, allowing for a more prosperous future with cheaper goods and shorter working hours.