Introduction
The current wave of automation, spurred by developments in artificial intelligence (AI), has been described as the second machine age (Brynjolfsson and McAfee, 2014) and the fourth industrial revolution (Schwab, 2017). One important part of this new era of smart machines is large-scale automation, not only of industrial production but also, and more importantly, of our everyday lives. Smart devices and the internet of things are supposed to make our lives, including our homes, smoother and more efficient. The historical descriptions of these tremendous changes often depict a linear development from steam-powered industrialisation and mass production (which also includes the invention of the railway and mass transportation of the first machine age) to large-scale digitalisation with the help of computers that is often depicted as replacing cognitive power. As Brynjolfsson and McAfee argue, what steam power was for the industrial age, the computer is for the second machine age.
This chapter aims to critique histories of automation that draw a picture of technological development as a teleological movement from industrial automation to âsmartâ machines, moving from the automation of manual tasks to automating cognitive labour. Instead, we demonstrate that technological innovation is never straightforward but characterised by failures and dead ends as well as specific choices that are anchored in the social and political contexts rather than a natural evolution towards the âbestâ technological solutions. Drawing on visualisations of automation in Swedish mainstream press since the 1950s, we focus on critical junctures of automation â such moments where it becomes apparent that automation develops into a different direction than initially imagined. By drawing on these materials, we emphasise the importance of mundane ways of imagining technological change as a way of meaning-making.
Material and analytical approach
The material on which the chapter is based is gathered from one of Swedenâs largest press clipping archives at the Sigtuna foundation. All newspaper clippings, which have been archived since the beginning of the 1900s until 2000, are sorted chronologically in envelopes by topic. The clippings have been collected from all the Swedish newspapers, as well as the largest newspaper in the neighbouring countries such as Finland, Denmark and Norway, including evening papers such as Aftonbladet and Kvällsposten, as well as quality national newspapers such as Dagens nyheter and Svenska dagbladet. As Johan Jarlbrink (2010) has argued, the Sigtuna clipping archive aimed to capture a broad picture of contemporary debates and favoured longer articles and essays over short daily updates on local events. Accordingly, the archival material not only represents a very specific selection of news but also effectively illustrates broad discourses in a specific time period.
We systematically searched the catalogue that is organised around keywords to identify relevant articles and visualisations. The press clippings themselves are not digitised or directly searchable. Hence, we systematically worked through the collections of clippings labelled âsocial issuesâ (including work environment and working time, unemployment issues and leisure), âtechnologyâ (including rationalisation of domestic work, technology from a social point of view, automation, operations analysis, electronics, history of technology, technical research, cybernetics and computing), âcrafts and industryâ (including computer industry and graphic industry) and âeconomicsâ (including machine culture, rationalisation, time studies, standardisation and technocracy) to identify relevant images. After having identified a set of relevant images, we returned to the digital archive of Swedish newspapers hosted by the Royal Library to extend the search with keywords based on our first findings in the clipping archive.
The advantage of working with this analogue clipping archive is that the search is broader and less dependent on changing terminologies. As digital development is characterised by its constant emphasis on being new and revolutionary, it follows that a key way to describe technology as new is through the way it is named. Hence, throughout history, digital media technologies have been constantly re-conceptualised. Consequently, as digital media technologies evolve, they often change names to illustrate their ânewnessâ and to distance themselves from older technologies. Terms such as automation, robotics, electronic brains, cybernetics, computers, mathematical machines, electronic data processing (EDP or ADP), AI, home computer, personal computer (PC), information technology (IT), information and communication technology (ICT) and algorithms partly denote different technologies in different time periods. However, they also overlap both technically and conceptually. By systematically reviewing press clippings, we aim to make emerging ways of thinking about and making sense of the computerisation and hence automation of society visible. We were, hence, exposed to far more articles than we would have been with keyword-based search in a database of newspapers.
The representations that we zoom in on here are emerging around two critical junctures: namely, the first automation backlash in the late 1960s and then the increased critique of data-based automation in the 1970s and 1980s. The two critical junctures in the automation of everyday life since the 1950s in Sweden illustrate how technological development is always also based on âpaths not takenâ. Automation in the 1950s was imagined as all-encompassing and concerning not only industrial but also, and in particular, the automation of household chores and everyday life. However, by the beginning of the 1960s, doubts emerged, and articles such as âAutoma-tion â is it really happening?â discussed the extent to which the dream of automation had, in fact, been realised. It also became increasingly clear that decisions to automate certain tasks rather than others were deeply entangled with power relations. Instead of a large-scale automation of the industrial sector, automation happened to a large extent at offices, involving tasks that were conducted by employees with a much weaker union representation than industrial workers at that time. The second critical juncture is an emerging discourse of surveillance and privacy that developed in the late 1970s and early 1980s. Instead of âdreaming bigâ in terms of technological change, the popular debate was increasingly dominated by automation anxiety (see also Bassett and Roberts, 2019). Automation was linked not so much to job losses but to questions of integrity and privacy that emerge with information gathering.
Taking these moments and controversies in the automation debate as a starting point, we focus our analysis on mundane ways of imagining technological development: namely, the visual depiction of automation in the popular press. In doing so, we approach technological development by analysing processes of representation as forms of meaning-making. The meaning-making around technology is related not only to how we imagine technology to work with and for us but also to which aspects we ignore in order to develop mundane, workable understandings of complex systems (Star and Bowker, 1999; Gitelman, 2008; Peters, 2015). Robert Willim (2017) has defined this process as mundanisation. With reference to domestication theory (Morley and Silverstone, 1990; Silverstone, 2005) that has conceptualised the integration of new technologies in our everyday lives â the taming of new technology â Willim argues that in order to be able to establish mundane uses of technologies, we have to highlight certain aspects while we forget and ignore others of these complex systems. As such, various lay theories about how technology works become important. We argue that popular media play an important role in conveying such theories: for example, by catering for general audiences rather than experts. We further argue that struggles and conflicting definitions around the meaning of certain technologies render technology and its process of mundanisation visible.
Imaginaries and what they teach us
1950s
In 1955, the Swedish Social Democratic Party and The Swedish Trade Union Confederation organised a now famous conference titled âTechnology and tomorrowâs societyâ, where it was argued that, in the future, computers will not only ârelieve the human workforce from heavy and monotonous muscle work, but also from tiring activities in the brain and nervous systemâ (Velander, 1956: 63). At this conference, the origin of automation is attributed to one of the directors of the Ford Motor Company in America, who during a meeting banged his fists on the table and exclaimed: âWe need more automation!â At the Swedish conference, it was speculated whether âautomationâ was a merging of the words automatic and production or perhaps just a mispronunciation of the word automatisation (Velander, 1956). Soon after that, the term âautoma-tionâ was further explained as an automatic process controlled by an âelectronic brainâ (i.e. computer) and quickly became a viral buzzword of the 1950s (Carlsson, 1999).
The mid-1950s is also the time period where the debate of automation exploded in Swedish popular media (Blomkvist, 1999). Here, it needs to be noted that the Swedish reformist labour movement has had a decisive influence on social development as the Social Democratic Party had uninterrupted government power for 40 years. The conference received a great deal of media attention, and the presentations at the conference were published in book form a year later. This means that early Swedish computer policy is to a large extent also the policy of the reformist labour movement, with a focus on controlling computerisations within the framework of the welfare state (Rahm, 2021). However, the debates in Sweden are not unique; in most countries in the industrial west, the media debate on automation and its potential for societal change began in the mid-1950s.
Automation is, at this point, imagined as revolutionising both the industry and the home, as it was predicted that robots will take over all human work in these domains. As an example, The British Department of Scientific and Industrial Research (1956) proposed that automation would take over virtually all jobs in industry. Despite this, there was also a strong belief that automation would not make workers redundant but rather improve working conditions (Dobinson, 1957). The automated future was expected to increase wealth, reduce the workload, create more leisure for all and thus increase the wellbeing of everyone. As Diebold (1958: 43) stressed in the journal Computers and Automation, âToday we are leaving the pushbutton age and entering an age when the buttons push themselvesâ.
Figure 1.1 shows an example of how automation was depicted in the 1950s. The caption reads: âthis is what it will look like when the administration is affected by automationâ. In a humorous way, automation is presented as a technology for the state apparatus, illustrating how machines were imagined as being in the service of governing institutions, citizens and even humanity at large. Basically, automation is visualised as a rationalisation of existing societal functions, while also remaining quite âsafeâ in terms of effects on society. You could argue that such depictions negotiate the function of technology in society by showing its positive and amusing side effects. The underlying plot can be seen as a vision of a âpush-button logicâ affecting households, industry and stately governance, where previously complex procedures are n...