Menu

Present Futures: Automation and the Politics of Anticipation

As an enduring contemporary concern, automation exists as much in the imagination as it does in practice. Yet this does not mean our imaginaries of automation should not be taken seriously. Particular actors have significant interest and investment in them, owing variously to the pursuit of money, influence and power. The circulation of ideas around automation establish what Foucault called ‘regimes of truth’, which Paul Harrison usefully identifies as “the very resources that allow us to agree or disagree”.

In the first issue of Futures of Work Andrew Sturdy and Glenn Morgan articulated how consultancies wield ‘thought leadership’ on the issue of automation. The management consultancy report constitutes the contemporary stand-out genre through which automation is imagined.The possible ‘futures of work’ put forward in this genre often play out in relation to perceived risks, with a proportion of workforce or sector defined as being at risk of displacement or replacement by automation. There is no mitigation offered, merely adaptation and readiness for which the management consultancy can help you prepare. These reports produce a ‘common sense’ understanding of jobs lost and ‘humans’ set aside on the scrapheap of history. In this way, as Sturdy and Morgan argue, consultancies move mere ‘matters of concern’ into the realm of certainty.

A politics of anticipation 

Articulations of possible futures of work shape discourse less out of the accuracy of the visions they project than the confidence placed in them. This shaping of discourse constitutes a ‘politics of anticipation‘. To paraphrase Bismarck, it is an art of conditioning the possible. More specifically, following Barbara Adam and Chris Groves we might argue that of the possible is conditioned by positioning the future in relation to the present in two particular ways. In one, “present futures” function as predictions and stand separate from the ‘now’. In another, “future presents” function as mechanisms to bring desirable outcomes into our current present(s).

“Future presents” are efforts to bring a ‘future’ into ‘the present’ to make it happen. “Present futures” are a way to hold a future at a distance in order to form the idea of a threatening risk or an attractive utopian goal.

These apparently opposing positions can be in play at the same time. They attain agency in enacting the worlds that are predicted, becoming self-fulfilling prophecies. In the view of the consultants espousing the opportunities and risks of automation, for instance, jobs are certainly going to be lost to automation processes. Therefore it follows that the companies they advise and influence should invest in automation strategies and technologies themselves, producing the job losses foreseen.

A proximate future? 

Predictions and projections about automation and futures of work always gesture towards a more-or-less proximate future. The proximity of this future horizon offers a semblance of credibility to the assertion that things will happen. This also lends urgency to what is anticipated, and with that can come pushback. As Mike Michael argues, “a near future can warrant swift action, but it can also attract the accusation that it is no more than opportunism on the part of the actor who gains from some sort of ‘scare’ or other”. Even so, ‘scares’ about job losses to automation seem to periodically gain traction, revitalising the newly commensensical regimes of truth on which they thrive.

A simple search of British national newspapers reveals how alarm spreads about job losses due to automation. For example, a simple news search for ‘automation’ and ‘robots’ reveals that, between September 2017 and August 2018, the Guardian published at least 22 articles about potential job loss within the next fifteen years. The majority of these are based, in whole or in part, upon either marketing materials for those selling the technologies concerned or on management consultancy reports. Eight of the pieces reported, or responded to, the findings of the PWC “UK Economic Outlook” reports of 2017 and 2018. Both featured chapters concerning the ‘risks’ of AI and automation. Similar reports by Deloitte andFuture Advocacy also feature as sources for national broadcasters and newspaper pieces covering the issue.

We can further unpick how these ideas about jobs displacement/replacement take hold and travel by following the references and sources used.

In an article on the Guardian website on 17th October 2017, economics editor Larry Elliot claims that automation “will affect one in five jobs” and that “workers in shadow chancellor John McDonnell’s constituency [Hayes and Harlington] face [the] highest risk of being replaced by robots”.

In this instance, certainty is enacted in forging a “present future” around the risk of wide-scale technological unemployment. But a “future present” is also laid out by specifying a context and a particular scale for the risk around McDonnell’s constituency of Hayes and Harlington.

The article reiterates the findings of a report by Future Advocacy, a centre-left think-tank, mapping the relative risks by sector for unemployment caused by automation in the UK. This is presented with some powerful choropleth maps visualising the relative risks. The report achieves this geographical granularity by using ONS data of employment per sector at the constituency level and applying national-level relative risk statistics for job losses due to automation produced by PWC in their Economic Outlook Report March 2017.

Between the article and two reports we witness two transformational steps in the movement of the narrative. “May” – a probability – becomes “will”, a certainty. This certainty is contextualised with a leap from the more abstract national scale to the more concrete local scale.

An issue with the transformation of data to tell a more compelling story at another scale with the same level of certainty is that, as geographers such as Ray Hudson, Doreen Massey and Michael Storper have argued, the distribution of labour does not translate as easily between scales as this assumes. This poses a potential issue for attempts to map risks derived at a national economic level onto the more granular level of individual parliamentary constituencies.

Regions, argues Storper, are where the action is. Factory closures owing to rationalisation of labour costs might appear uniformly distributed at the national level. But the impacts will be felt in differentiated ways at a regional level. We might call this a ‘modified area problem’ – data derived at one scale does not necessarily work at another scale. The form of spatial imagination enacted in these predictions matters.

Role or task automation? 

A further step can be charted from management consultants to academics. In the PWC reports, as well as reports by Deloitte, there are frequent references to other sources upon which their own analyses are based. These are principally two papers: the 2016 OECD social employment and migration working paper no. 189 by Arntz, Gregory and Zierahn and a 2013 paper, “The Future of Employment: How susceptable are jobs to computerisation?“, by two Oxford academics Frey and Osborne.

These papers present conflicting visions of the future of work that help unravel the certainty of the “present future” portrayed in the aforementioned Guardian article and throw into doubt the idea of whole roles being displaced or replaced by automation. Instead, these academic analyses focus on ‘tasks’ being automated because, as Arntz and his co-authors argue, focusing on jobs “might lead to an overestimation of job automatibility, as occupations labelled as high-risk occupations often still contain a substantial share of tasks that are hard to automate”.

All this goes to show the ways in which ideas around a particular future of work travel, the work the ideas do themselves and the work that their translation enacts upon them. Likewise, these all serve to shape how publics variously negotiate those ideas.

In part this is a story about the public communication of research, wherein nuance is lost in favour of a punchy headline. But there is much more to how a politics of anticipation plays out in the negotiations and translations of the present futures of work. The formulation of a future as the future, as argued in the first editorial of Futures of Work, does political work – it enables certain forms of powerful rhetoric, for example around a fourth industrial revolution. Rhetorically constructed as the “present future”, it possesses a discursive power over what is considered ‘common sense’ and elides the persistent contradictions of capitalism. Meanwhile, it allows management consultancies and others to propose the proximity of a “future present” as a problem for which there is a solution to be sold, as Sturdy and Morgan argue.

All too often arguments around automation and futures of work accept narratives of displacement or replacement of jobs as their premise. This renders them common sense, rather than calling them into question. How we construct and receive ‘truths’ about possible futures of work is political and demands political action. We must pay attention to how particular futures become ‘facts’ and how well those ‘facts’ travel and take on a life of their own.

Sam Kinsley is a Senior Lecturer in Human Geography at the University of Exeter

Image Credit:  Phonlamai Photo/Shutterstock