New report shows lessons from the past can help societies navigate an uncertain employment future

13 June 2019

Adobe Stock 227994414
A new GPS report from the Oxford Martin School and Citi analyses the future of labour markets in a world on the brink of being transformed by advances in artificial intelligence, machine learning and automation.

AI and machine learning promise not just to alter the structure of our economy but also to reshape our social fabric more broadly, in no small part due to their widespread impact on employment. History has shown technological progress creates as well as destroys jobs, a process that can be very difficult for large parts of society for a very long time. The new report - Technology and Employment v4.0: Navigating the Future of Work – sets out what we can learn about the implementation of such technologies from past periods of disruption, such as the Industrial Revolution in the UK, and what these lessons might mean for future periods of rapid technological change.

The report opens with insights from Dr Carl Benedikt Frey, co-director of the Oxford Martin Programme on Technology and Employment.

His findings include:

· From driverless cars to facial recognition technology, AI and machine learning applications are improving rapidly. As with earlier General Purpose Technologies (GPTs) such as steam power, electricity and computing, these new technologies have the potential to reshape the structure of the entire economy. But they are also likely to follow a similar path when it comes to productivity, where gains are likely to be delayed while companies rethink their operations in order to take full advantage of new technology.

· Integrating automation is a complex process for companies, who must consider the cost of replacing old technologies and processes, the availability of markets, financing for investment in new technologies and skills, legislative hurdles, worker resistance and public opinion. These factors are as relevant for AI and machine learning as they have been for other technologies.

· In times of worker-replacing technological change, the flipside of the cheapening of goods is significant parts of the workforce may suffer in their capacity as producers in the labour market. As in the period immediately after the Industrial Revolution, advances in automation have caused a similar hollowing out of middle-income jobs, with the gains from growth shifting from the bottom to the top of the income distribution and from labour to owners of capital.

· The continuation of current trends in the labour market could lead to further regional divergence and increased polarisation in society and politics. AI is likely to create more highly skilled jobs while replacing unskilled ones, leading to jobs booms in cities that are thriving tech centres, while negatively impacting areas where people are more likely to be employed in routine manufacturing jobs.

Dr Frey said: “The hype might seem exaggerated for now, but over the next decades AI will have pervasive effects on productivity. There are many bottlenecks to technology adoption, so the mere existence of a new technology is not sufficient for it to find widespread use.

“Current productivity growth is therefore an inherently bad predictor of future productivity growth. In fact, in the early stages of development new technologies may even reduce productivity growth. Take, for example, the development of autonomous vehicles. The Brookings Institution estimates between 2014 and 2017 investments in autonomous driving amounted to roughly $80 billion, with very few first-case uses so far.”

The report also draws on Dr Frey’s new book, The Technology Trap: Capital, Labor, and Power in the Age of Automation (Princeton University Press).

He added: “Innovation itself is not what drives productivity growth. For productivity to grow technological innovations must find widespread use. The spread of every technology is a decision, and if the groups that stand to lose out from the technology are sufficiently powerful, they might succeed in blocking it. Progress is evidently not inevitable and for parts of the population it might not even be desirable.

“Technological change tends to come with creative destruction in employment, meaning there will always be both winners and losers, at least in the short run.”

· Download ‘Technology and Employment v4.0: Navigating the Future of Work’

· Find out more about the Oxford Martin Programme on Technology and Employment

· Buy The Technology Trap: Capital, Labor, and Power in the Age of Automation