Situating and expanding the scope of dispositions towards automation

Abstract

The four generous commentaries to our article, ‘Dispositions towards automation’, have truly enriched and extended our thinking on ‘dispositions’ as a way of challenging binary readings of automation (i.e. either towards capitulation or adaptation). Two of our interlocutors have encouraged us to think more situationally about the way dispositions develop through user experience and encounter, while the other two welcome a further expansion of our ideas to include infrastructural and other non-human or environmental agencies. In this response, we clarify our original curiosities about capital's affective projects with respect to automation, as well as its attempts at steering dispositions in order to mitigate, precisely, the indeterminacy of labour and user ‘reactions’. At the same time, we discuss how the four provocations have inspired us to re-assemble these tendencies and sentiments more plurally, to incorporate a series of overlapping affects, capitalist plots, shifting design values, detours, exigencies, digital glitches, and material interjections. We are impressed by the holism of perspectives presented in this dialogue, and how the discussion has only reinforced ideas about the fragility of automation's relations.

Introduction

Since the publication of our original article, a couple of headlines on automation in March 2023 had grabbed our attention. The first involved an open letter signed by over 1100 individuals, including Elon Musk, Steve Wozniak, and Victoria Krakovna, who questioned, somewhat self-incriminatingly, whether non-human minds should be developed in ways ‘that might eventually outnumber, outsmart, obsolete, and replace us’ (Rudnitsky and Metz, 2023). Expressing another state of alarm, the second news story conveyed, more locally, the frustration of delayed travellers at Changi Airport – one of the airports we mentioned in our article – who were beset by a system-wide glitch in the air hub's automated immigration clearance facility. Unable to rely on the usually efficient, face-recognising smart gates to process those leaving Singapore, passengers joined ‘super long queue[s]’ for ‘manned immigration clearance counters’ amid an air of disenchantment (Channelnewsasia, 2023).

Notwithstanding the differences in scale and context, this pair of headlines captures the paradox of two realities currently circulating in debates on automation. On the one hand, there are broad references to – and often critiques of – large political projects in artificial intelligence (AI), machine learning, virtual augmentation, and robotics development that animate the design scene (Graham et al., 2022; Macrorie et al., 2021). On the other hand, there are also the more mundane accounts of ‘actually-existing’ (dis)uses of, encounters with, and resistances against automation that challenge technological dreams in real life, albeit against which those dreams can be felt and judged (Bissell, 2021; Richardson, 2018).

Looking at the four thoughtful replies from our commentators, our desire is to revisit and ponder these two positionalities again and ask where our paper stands. Do the dispositions we surface risk falling into the trap of either position? Do they encourage further reflections on automative providence and use? Might they serve as starting points to contemplate how dispositions develop and dissolve in indeterminate ways? What is the role of the positive and technophilic feelings that we countenanced? We believe these questions are important not only to this set of current papers but also to other dialogues parsing between the construction and apprehension of automation.

Situation

We begin with Zhao's (2024) much-welcome reminder on ‘the legitimacy and necessity of situating a disposition in place’. Drawing on a postphenomenological perspective, he argues that technology/automation must always be contextualised within its interactions, such that any disposition it instigates must, at best, be treated as momentary, and always ‘a mirror of its user and the place it is situated’. For Zhao, such a stance makes room for more complex encounters with automation that does not simply result in, as it were, automatic adaptations to technology, but, possibly, its rejection, amnesia, neglect, and a range of other more ambivalent emotions.

Similarly, Bissell (2024) echoes this thesis, pointing out that any ‘reaction’ – as too strong a word as it is – to automated technologies, is never sustained, but ‘highly likely to be subtle and fleeting at most’. Positing that many of the aeromobile dispositions identified by us ‘might actually be pretty weak or even non-existent’, he suggests the need to trace these (longer-term) dispositions back to their ‘choppier short-duration’ constitution by situated ‘encounters’. For Bissell, paying attention to such banal details would allow the agency to be returned to labour, passengers, and other users of automation, who, rather than ‘anaesthetised’ and ‘passive marionettes’, are imperfectly affected subjects with the potential for transformative and open-ended dispositions.

We are grateful for these thoughtful cues and do very much agree with Bissell and Zhao on the need to consider more fluid technological affectations at the subjective level. We offered the skein of dispositions mentioned perhaps with not enough assertion of their tentativeness, their fragility, plurality, and even weak force, ‘relative to the dispositions that are more front and centre for people’, as Bissell writes. And we wonder if those dispositions can interfere, perhaps conflict and support those front and centre ones?

We might add the idea that such situatedness stretches the other way to apply to how capital operates as well. To return to the introduction of this response, although automation's drivers (e.g. inventors, designers, developers, money-makers) may be inclined to construct dispositions in particular ways, they rarely manage to achieve the perfect ‘infrastructural ideal’ (Manderscheid, 2014). Instead, they expect these dispositions – and hence their operation of power – to be uncertain, tentative, and fragile. This is also why automation's drivers often have had to conscientiously change course, invent new discourses, redesign their products, and even withdraw their initiatives, in order to appease and respond to varying uptakes of their products. In other words, just as how users and workers encounter technology situationally, so too do these agents of power, who must adapt to elicit desired reactions, enthrall, seduce, and protect their interests through revamped ‘technologies… that are made to interpellate (otherwise-reluctant) subjects’ (Lin et al., 2024).

This schema brings us back to the crux of our article, which is to try to (re)imagine dispositions as somehow (partly) manufactured, elicited, steered, and designed to ‘meet’ technological subjects halfway. This is not to say that capital has complete dominance over the process of defining automation (it does not) but it is also not quite an admittance that capital relinquishes control of encounters once the initial design phase is complete (design is never complete). In other words, we are seeking to emphasise the fact that capital is continually trying to foster certain sentiments and helping them settle in the way that it wants through affective ‘meeting points’. As Bissell and Zhao rightly point out, any outcome is indeed at the mercy of situations that may take unexpected directions, but, with a bit of luck, some of these crafted dispositions might gather traction and take root.

Expansion

Here, we want to turn to the contributions of Ash and Fraser, who perhaps provide another way to square the binary ‘realities’ of grand automative projects versus everyday encounters. In Fraser's (2024) reading, she begins with a commensurate sensitivity to the importance of actual experience, but not without a simultaneous interest in ‘interfaces, procedurality, and mobile media’ – in other words, infrastructure – that frame and precede real-world interactions. As she suggests, automated technologies may not be the direct pathways to dispositions like gamification, but they nonetheless serve as subtle mechanisms, hidden abstractions, and concealed enablers ‘of the growing intercession of the computational, machinic, and robotic into processes that were once fully human’. In short, while never fully affective, each technical infrastructure has at least an intent to influence the structure of feeling towards it and towards subjective experiences.

More ambitiously, Ash's (2024) piece alerts us to the possibility of further broadening the focus of dispositions beyond the human, or even automation itself, to more expansively include other non-human entities that are normally out of the picture. Inquisitively, he probes whether it is possible ‘that all entities, technical or otherwise also have a disposition and so are disposed to one another in different ways’. Here, a putative ‘environmental’ perspective is assumed, and at stake are not just closed technological assemblages but compositionally fluid milieus of actors – from algorithms to speeds to various materialities that come into play – that each can shift affective tendencies towards technology. Making such a leap exponentially pluralises (Bissell, 2024) the origins of dispositions, going beyond capital, subjects, technological designers, and technical infrastructures to co-opt a whole host of other ‘environmental’ (f)actors that are impossible to pre-determine before they come into view. We also see some possible parallels with Amoore's (2021) recent attention to the ‘deep border’. Amoore explores the unlikely resonances between the layered and complex operations of deep machine learning, with the expansion of the state's reach into the complex and multi-layered digital and data entanglements of everyday life. It is not only that border governance practices may make use of increasingly complex and sophisticated forms of machine learning but that the shape and practice of the border seems increasingly disposed to follow these logics, and indeed to attend to the very ‘leanings, inclinations, and propensities’ garnered from inferences over the attributes of populations.

In many ways, such an expansionist viewpoint holds immense appeal to us, for it further undermines the ‘simple “hope or fear” binary readings of technology’ (Ash, 2024). Indeed, to revisit the headlines at the start of this response a second time, the actors implicated in these stories go far beyond just AI technology and aggrieved passengers/labour. They also include leading system developers in the US wanting to throttle down the pace of innovation, misgivings about AI's rapid progress there (or simply fear of competition), digital faults at an airport halfway round the world, vacated immigration desks thanks to automation, and tight flight schedules that could not materially accommodate delays and ‘environmental’ events (whether of the weather or automative malfunction). These twin sets of interactions do not happen in neat sequences, neither are they even cognisant of each other. Would these disparate actors ever enter into conversation with one another; if not now, would they relate to each other later on; might they be entangled with other happenings before and after them? These questions are hugely complex, and if anything, remind us of the innumerable overlaps of affects, capitalist plots, shifting design values, detours, exigencies, digital glitches, and material interjections that populate what we know of as ‘automation’.

Of course, expanding the scope of automation assemblages and their dispositional sources in this manner can greatly complicate research and its methodologies. It requires a commitment to attune to the myriad agencies that impress upon automated systems, as well as a pursuit of flexible modes of engagement to capture them in the past, present, and future. As one of us has argued elsewhere, technoscience unfolds ‘as a dynamic skein of activities and sequences, rather than a time-limited snapshot of fixed relations’, having ‘no set paths as to what might happen, or how things might accrete’ (Lin, 2023: 3). This is exactly the kind of ambiguity and fraught relations that we had wished – albeit, it appears now, in very small part – to foreground through the idea of dispositions in design. Perhaps in different ways, this fragility is also what our commentators have insisted.

Further reflections

We started out this dialogue by sketching several ‘dispositions towards automation’ that capital might be formulating to ease transitions towards automated futures after its economic agendas. Meant neither to be exhaustive nor set-in-stone, the five dispositions we originally proposed set out to destabilise conceptualisations of automation as either an unstoppable force that threatens to displace labour or as a static infrastructure its users slowly chisel away at in their daily practice. Indeed, their very proposition signals our concern for the process by which capital tries to stabilise its rule – through affective manipulation and acquiescence – rather than the absolute success/failure of these measures. For us, these dispositions testify more to capital's vulnerability than its strength, for it is seen to constantly persuade, cajole, entertain, and even sway those whom it seeks to influence.

Our commentators have given us – as well as those who are interested in the subject – several provocative prompts that we believe are able to take this debate in new and exciting directions. They have reminded us of the need to be careful about taking these dispositions too far, for they may not ever take off or actually exist in practical situations (without prejudice to capital's further attempts to re-engage, of course). They have also encouraged us to expand these dispositions beyond human concerns to pay attention to infrastructural, mediative, and other non-technical, non-human dispositions that can equally shape and complicate affective relations. We are heartened by the keenness our commentators have shown in challenging our ideas and making them more refined and holistic. Like Fraser, we, too, hope for further conversations and collaborations on this subject with our commentators and beyond – and indeed many more publications that are able to draw out the richness of what we have together started to confront.