This special issue of New Formations explores contemporary cultural anxiety about new forms of automation. Automation anxiety, in the widest sense, could perhaps be applied to cultural debates and concerns about machines from the creation of factories at the beginning of the industrial revolution to the implementation of fully-automated manufacturing in the 1980s. Coinage of the contemporary term automation, however, is usually attributed to Ford vice-president Delmar S. Harder in 1948.1 Harder’s neologism became the topic of widespread debate in the 1950s and 1960s, taking over from ‘mechanisation’ as the object of concern (and of utopian dreams) regarding the machinic substitution of human labour. Automation in the postwar era meant something more systematic and extensive than mechanisation: the creation of closed feedback loops to control whole production processes, as well as the addition of electronics and digital computers to mechanical solutions. Most importantly, the term automation was applied not just to actually existing developments. When Harder said that ‘what we need is more automation’ he was referring to relatively modest improvements to the way worked materials were transferred from process to process on Ford’s production line (p149). But the term was quickly taken up to refer to future technological developments, developments which promised– or threatened– dramatic changes in both the quantity and quality of tasks that machines could accomplish. This sense of automation as something to come, an essentially speculative dimension of technological progress became inseparable from a certain anxiety about the imagined future and its implications for work and leisure.
From self-driving cars, through high-frequency trading to military drones and organised swarms of shelf-stacking robots our era is seemingly characterised by a new wave of automation. The wider topic of automation is a pressing subject with various existing academic responses, particularly in relation to the future of work, the automation of warfighting, and the algorithmic management of social existence.2 The focus of this special issue is to address, as a topic in its own right, the cultural and social anxiety generated by these new forms of computational automation. Automation anxiety here can equally imply, for us, automation fever: instrumental or utopian demands that ‘what we need is more automation’.
The current ‘rise of the machines’ is characterised by the replacement of complex cognitive tasks and human decision-making by algorithms, machine learning and other computational techniques. Automation anxiety concerning these developments is evident in many contemporary public debates and political interventions. In their 2013 working paper ‘The Future of Employment: how susceptible are jobs to computerisation’, Carl Benedikt Frey [End Page 5] and Michael A. Osborne tried to predict the impact of recent developments in machine learning on the automation of different occupations.3 Their analysis arrived at the startling conclusion that 47 per cent of current jobs in the US could be performed by computers within the next two decades. Frey and Osborne’s work was much quoted and widely disseminated publicly. In September 2015 the BBC applied their findings to UK occupational data and created a web page which provided searchable estimates of its readers’ jobs being automated. According to this web page, train and tram drivers had a 68 per cent chance of having their role performed by computers in the next two decades. On the other hand, the chance of ‘higher education teaching professionals’ being replaced was only 3 per cent.
Frey and Osborne’s analysis relied on a workshop held at Oxford University Engineering Sciences Department where machine learning researchers were asked to classify the likelihood of occupations being automatable in the near future. This classification was then used to build a predictive model of the likely attributes of automatable jobs and applied it to information about job roles in O*NET, an occupational database originally created for the US Department of Labor. Both elements of Frey and Osborne’s approach have come under sustained critique since 2013. As Melanie Arntz, Terry Gregory and Ulrich Zierahn argued in a working paper for the OECD in 2016, Frey and Osborne rely on engineering experts to judge the automatability of occupations, despite good evidence to suggest that such experts tend...