Automation and environmental dispositions

Abstract

To think through the dispositions automated systems attempt to generate, it is key to understand how automated systems relate to the environments in which they operate. Developing Lin et al.’s important arguments around dispositions towards automation, this short response suggests the concept of disposition can be broadened to think about non-human dispositions more widely. To do this, the response forwards the notion of environmental disposition; a kind of fundamental state that all entities find themselves in, whereby entities are compelled to enter a more or less provisional position or grounding in an environment or milieu.

Introduction

In ‘Dispositions towards automation’, Lin et al. (2024) offer a wonderfully rich conceptual narrative of the shifting geographies of automation. Two of the key points of the article are particularly important and worth emphasising. First, it is important to challenge simple ‘hope or fear’ binary readings of technology, which seems to be the dominant way that ‘new’ technologies, such as automated systems, come to be discussed in popular discourse (Kinsley, 2012). Avoiding binary readings of technologies such as automation, which Lin et al. argue often fall into narratives of ‘capitulation and adaptation’, requires other categories or modes of analysis of these technologies. Second, it is important to go beyond naming general processes (such as automation) and to develop vocabularies for understanding distinct differentiated tendencies and systems within a particular domain of technological development and practice. In their paper, Lin et al. do this by offering the reader ‘dispositions towards automation’. These dispositions are less about what automated systems are, and more about how they can generate modes through which humans can relate to them. In this short response, I want to focus on what is at stake in Lin et al.'s definition of disposition and how the term might be broadened beyond its focus on human subjects to think more about the non-human entities involved in automated systems.

Environmental dispositions

Lin et al.'s account of disposition emerges from work on the sociology of habitus (Bourdieu, 2015) and the work of geographers such as Lorimer (2005) and Bissell (2021). Lin et al. define disposition through work on embodied practice and their definition is clear. Dispositions are ‘socially and culturally inflected and transmissible affective capacities to affect and be affected in each encounter and setting’ (Lin et al., 2024). In other words, dispositions could be understood as tendencies to act, think, work, or move that is an outcome of an encounter with another entity or person, which can, in turn, alter what that entity or person is or what they can do and vice versa. In utilising this definition, Lin et al. are careful to avoid thinking of dispositions as purely intentional forms of activity. Instead, they see dispositions as potentially passive as well as active, which can involve ‘the diminishment of capacities … acquiescence, loss or giving way’.

Lin et al. also draw upon assemblage theory (Anderson et al., 2012 ; Delanda, 2016; Deleuze and Guattari, 1987; Müller and Schurr, 2016) to emphasise that dispositions are not static or pregiven. In their words, ‘dispositions … [are] … constituted in and through … assemblages’, where ‘a disposition … is … gradually shaped through or transformed by multiple events’. Here, ‘particular dispositions are located … not then in or necessarily of the body or a single entity itself’ (Lin et al., 2024). Different dispositions towards automation can therefore be defined as ‘the in-process tendencies, or “points of meeting,” of human-technology assemblages which enable, creatively rework and resist automation’. Lin et al. discuss five of these dispositions, which they term ‘enchantment, aspiration, experimentation, gamification, and acquiescence’. The aim of naming these different modes is to emphasise the ‘conditionality’ of the ‘conflicting relations’ between humans and automated systems.

Linking practice and assemblage theory together to define disposition and dispositions to automation is interesting. With a focus on practice, the paper emphasises the fact that dispositions are primarily human, and it could be read that disposition is a one-way process. That is to say, dispositions are a matter of how humans are disposed by and towards something (in this case automated technologies). As Lin et al. put it, ‘automation, coupled with a cognisance of feelings towards it, can contribute to activating particular dispositions and subjectivities among technology's users affecting their capacities to affect and be affected’. At the same time, in suggesting that dispositions fundamentally emerge through events and are ‘not then in or necessarily of the body’, the paper points to the idea that dispositions not only affect humans and are not only a one-way process from technology to human. Nonetheless, the paper does tend to focus on dispositions as ultimately a matter of human concern. For instance, automated airport systems are designed to ‘orient passengers towards’ an ‘assortment of delightful dispositions’, and ride-hailing ‘[c]ompanies like Uber use gaming techniques that directly bear on the feelings and experiences of labouring bodies, most often in an attempt to encourage drivers to work longer and in ways that may be less beneficial to them’.

However, would it be possible to push the logic of an assemblage perspective to suggest that not only do automated systems generate dispositions, but that all entities, technical or otherwise also have a disposition and so are disposed to one another in different ways? Here it might be useful to think about disposition as not only a kind of affective capacity to respond to a situation. Instead, a researcher might begin by thinking about disposition as a kind of fundamental state that all entities find themselves in, whereby entities are compelled to enter a more or less provisional position or grounding in an environment or milieu. For example, even a simple non-human technical thing such as a brick wall settles into the ground after being built and attempts to come into some kind of equilibrium with its environment over time.

This expanded notion of disposition could result in a potentially altered approach to studying dispositions towards automation by working to think about environmental dispositions. Thinking about environmental dispositions would necessarily involve analysing the dispositions of the entities that make up the environment in which an automated system was operating (which would include all manner of humans and non-humans) and how those environmental dispositions inform the disposition that the automated system was trying to compel in the user. For example, Lin et al.'s discussion of Uber demonstrates how the programming of the Uber app operates through the disposition of gamification, which attempts to create mystification to lure people into both working as Uber drivers and becoming Uber passengers. Utilising a broadened notion of disposition, a researcher might also seek to understand how the gamified disposition to automation both draws upon and feeds into other environmental dispositions that make up the driver/passenger/city assemblage. For example, how does the Uber app dispose itself towards the Google Map APIs that it uses to calculate journey distances and how does the disposition generated in the relation between the app and API constrain and enable the disposition of gamification that Uber attempts to cultivate? Furthermore, how might the disposition between the Uber app and Google Map API dispose vehicles to using particular routes through the city? In turn, how might these layers of disposition dispose the driver to move through the roads as they work? Do the above dispositions affect the speed the driver travels? Do they affect how often the driver applies their brakes, or when the driver decides to turn their engine off? Subsequently, do these dispositions toward driving and the vehicle's disposition to the driving environment cause specific forms of wear on roads or potentially create hotspots of atmospheric pollution?

Here, dispositions towards automation become automated dispositions that work across and through human and non-human realms and stretch far beyond distinct boundaries of human labour or agency. In short, can Lin et al.'s important development of the concept of disposition be usefully expanded to understand how automated technologies are disposed to other automated systems as well as technologies and other humans and non-humans more broadly? From this perspective, technical objects (including automated systems) find themselves disposed to one another, as much as they are disposed to humans or other non-humans. In turn, dispositions towards automation can be understood as fundamentally linked to the broader environments and milieus in which automated systems operate.

Conclusion

In the conclusion to their paper, Lin et al. suggest that the concept of disposition helps ‘identify automation … as, foremost, cultural – and not just technical – form(ul)ations that are deeply intertwined with specific societal expectations, ambitions, and strategies unfolding and generated across space and time’. In this response, I have suggested that expanding the concept of disposition to think more about the specific dispositions of all entities allows a broader understanding of human-technical formations and automated assemblages in particular. Thinking about how automated technologies are disposed towards one another, other technologies, and non-humans in general, as much as they are disposed towards humans (and humans are to them), might help to widen future theoretical and empirical research into automated systems.

Of course, thinking about automated systems in terms of environmental dispositions raises many questions. The first and perhaps most important being one of identifying boundaries and domains of automated systems and seeking to understand what entities become disposed in relation to what automated systems and how. For example, which entities are exactly involved and disposed to one another when an automated driving system decides to brake or accelerate (Ash, 2017)? Inspired by Lin et al.'s work it is possible to argue that identifying these domains become less about what automation is or means, but more a matter of analysing how automated systems always already exist alongside environments and milieus of different bodies, technologies, and practices (Gabrys, 2014, 2015; McLean, 2019). Here dispositions towards automation becomes a matter of how bodies, technologies, and practices are disposed (or maybe even predisposed) to one another in ways that alter what these bodies, technologies, and practices are and what they can do within a particular domain. Lin et al.'s development of the concept of dispositions towards automation thus provides an important contribution to debates on automation, automated systems, and their multiple geographies.