Automation: ‘a foreboding, frightening word’
Fears about advances in automation leading to mass unemployment are nothing new. In nearly every decade of the twentieth century, in fact, there have been fears about what John Maynard Keynes called ‘technological unemployment’. In the 1920s, the New York Times balked at the hysteria for ‘machines, machines, machines!’ and asked if man would be “enslaved by the instruments he creates to free him from bondage”.
Thirty years later, when advances in automatic machine control created the “uncanny spectacle” of “machines without human beings”, the United Auto Workers union (UAW) described automation as “a foreboding, frightening word [...] a word to strike terror in any human heart, especially if that human heart beats in a wage earner's chest.” By the late 1970s, Horizon was warning that microprocessors would be the reason why Britain’s children would “grow up without jobs to go to.” In all of these cases, however, the predicted mass ‘technological unemployment’ failed to materialise.
Will this time be different?
In keeping with this trend, the last few years have witnessed almost obsessive media coverage of the threat of artificial intelligence to employment. Will new technologies ranging from Chinese industrial robots to advances in machine learning finally lead to mass unemployment on the scale predicted over the past century?
This question was one of the motivations for Nesta’s joint research project aimed at identifying how social, economic, and technological drivers of change will affect the jobs and skills we’ll need in 15 years’ time. The first phase of the study has focused on how historical trends relate to the evolving composition of the workforce. Alongside this research, Nesta is also preparing to host FutureFest, a two-day festival (September 17-18) during which attendees and speakers will gather to debate questions related to how we will work, play, love, and thrive in the future. As part of the first of these themes I will deliver a talk on ‘the history of the future of automation’, complete with clips from films, documentaries, and newsreels from the past century. As we move forward with both projects, now would seem to be a good time to explore past debates about automation in the workplace and to examine the fears that have so often surrounded the introduction of innovative technologies.
By 1931, Modern Mechanics and Inventions was already referring to these fears as an “old cry”
Perhaps surprisingly, drawing attention to the recurrent nature of fears about automation has a history nearly as long as the fears themselves. By 1931, Modern Mechanics and Inventions was already referring to these fears as an “old cry”. Indeed, many generations have been aware that their ancestors have repeatedly made predictions of mass unemployment only to see the economy and labour market adapt. Yet these same generations have convinced themselves that they are the ones living in the age when the exception finally proves the rule. Part of the reason for this seems to be that they have imagined themselves living through a period of unprecedented and rapid changes relative to the changes experienced by previous generations.
A good example of this can be found in the testimony delivered before Congress in 1955 by Walter Reuther, President of the UAW. Reuther began his testimony by expressing his union’s belief that America was “really standing on the threshold of a completely revolutionary change”. From the beginning of the Industrial Revolution through the 1940s, he informed the panel, the pace of automation had been gradual and relatively stable. As the world moved into the 1950s, however, advances in automatic control threatened to “substitute the thinking process on a mechanical basis for the thinking process which heretofore was done exclusively by the human mind”. This convinced Reuther and the UAW that the impact of automation on the economy and workforce over the next decade would be “much greater than the impact of the first phase of the Industrial Revolution”.
Fast forward two decades and commentators were instead writing that it was the seventies and eighties that would witness a period of mass unemployment and unprecedentedly rapid changes to the composition of the labour market. A 1978 BBC Horizon programme entitled, ‘Now the Chips are Down’ re-interpreted advances in automation during the 1950s as contiguous with the advances introduced in the two hundred years since the Industrial Revolution; it was the computer chip and computerized control, the programme explained, that represented a fundamental shift in the nature of automation. “Can we all live on […] the earnings of an elite band of 60,000 software engineers?”, asked Horizon. To do so would mean “condemning a generation to unemployment.”
The common claim that the digital revolution of the last 30 years has drastically altered the kinds of skills required in the workforce, can be misleading
Research from the OECD, however, has shown that while the direction of change in the labour market since 1950 has indeed been steadily away from farm and production jobs and toward professional and technical jobs, the pace of that change has been quite gradual. The common claim that the digital revolution of the last 30 years has drastically altered the kinds of skills required in the workforce, can be misleading.
When confronted with innovative technologies that appear to represent fundamental advances in automation, past generations have routinely overestimated the potential impact of these technologies on the composition of the labour market. Part of the reason for this is likely that these prognostications have not taken account of the myriad mitigating, non-technological factors at play. This idea was explored more fully in Nesta’s 2014 book on the promise of a robot economy edited by Stian Westlake, which starts with the statement: “How society uses new technologies is not a foregone conclusion. It depends on political decisions, cultural norms and economic choices as much as on the technologies themselves.”
...it seems that developing an understanding of the potential continuities in the labour market is as important as understanding the potential discontinuities.
These findings are encouraging for Nesta’s research into demand for jobs and skills in 2030, for it is clear that examining historical trends can prove valuable for studies of the future of the labour market. More importantly, it seems that developing an understanding of the potential continuities in the labour market is as important as understanding the potential discontinuities.
To join in this centuries-long debate, come along to my talk at FutureFest, ‘Apparently the robots are still coming’ (12pm on Saturday; 12.30pm on Sunday). And to learn more about the 2030 project, don’t miss Mark Griffith’s talk at 11.30am on Saturday.