Machine Habitus. Massimo Airoldi

of cultural content and filtering it in path-dependent, personalized ways (Beer 2013a; Morris 2015; Prey 2018; Fourcade and Johns 2020). Algorithmic systems have a role in the amplification of material and symbolic inequalities, as witnessed by automated forms of discrimination against the poor in the US (Eubanks 2018), or by the ubiquitous computational reinforcement of race and gender stereotypes (Noble 2018). Although they do not have a culturally connotated accent, like the French peasants discriminated against by the Parisian elite (Bourdieu 1991), chatbots and digital assistants may use different vocabularies and registers depending on their training and past communications. These machines are certainly different from human beings, but they perhaps contribute even more than we do to the ‘reproduction’ (Bourdieu and Passeron 1990) of an unequal, yet naturalized, social order.

      It is important to note that, according to Bourdieu, the historical reproduction of social inequalities and discriminations is not the deliberate outcome of a coherent apparatus of power – as it was for Marxist scholars of his time (Boudon and Bourricaud 2003: 376). Rather, the perpetuation of the social order is the aggregate result of myriad situated encounters between a habitus’ cultural dispositions and a field – that is, a ‘domain of social life that has its own rules of organization, generates a set of positions, and supports the practices associated with them’ (Calhoun et al. 2002: 262). Examples are the fields of cultural production (Bourdieu 1993) and consumption (Bourdieu 1984), as well as the education system, with its inner hierarchies and repressive institutions (Bourdieu and Passeron 1990). On the one side, ‘habitus contributes to constituting the field as a meaningful world’ and, on the other, ‘the field structures the habitus’ (Bourdieu and Wacquant 1992: 127) through its implicit rules and common sense – doxa, in Bourdieusian jargon. From this theoretical viewpoint, any form of domination, such as that working-class people (Bourdieu 1999) or women (Bourdieu 2001) are subjected to, can be seen as the subtle, naturalized outcome of pre-conscious power mechanisms rooted in culture.

      What if we extend Bourdieu’s inspiring ideas to the cold technical realm of algorithms? What if we start seeing machine learning systems as socialized agents carrying a machine habitus and recursively interacting with platform users within the techno-social fields of digital media, thus practically contributing to the social reproduction of inequalities and culture? This book is a journey through these largely unexplored theoretical landscapes, where two main kinds of agents – humans and machine learning systems – and their cultural ‘black boxes’ – habitus and machine habitus – jointly make society, while being made by it.

A closer, comprehensive look at the cultural dispositions of machine learning systems – their formation and translation into practice, their influence on human habitus and transformation over time, across social fields and domains of applications – can serve to advance our understandings of today’s techno-social world. Certainly, we already know a lot about it. Whole fields of research are dedicated to the study of algorithmic bias, human–computer interaction and machine discrimination. The purpose of this book is to restate the obvious in a sociologically less obvious fashion, deliberately designed to ‘transgress’ disciplinary borders, as suggested by Bourdieu himself (Bourdieu and Wacquant 1992: 149).

      The following chapters attempt to bridge insights from cultural sociology and computer science, AI research and Science and Technology Studies. Chapter 2 will help us understand the social genesis of the machine habitus, and how it must be distinguished from another type of culture in the code, present in any technological artefact; that is, the culture of its human creators, acting as a deus in machina.⁷ Chapter 3 will focus on the code in the culture, aiming to theorize the forms and effects of algorithmic practice, and how these concretize within the techno-social fields of digital platforms. Chapter 4 will then bring the culture in the code and code in the culture together in order to sketch a general theory of machine habitus in action. The concluding chapter will highlight the sociological relevance of mechanisms of techno-social reproduction and propose alternative ways to imagine, design and relate to socialized machines in real life.

Notes

      1  1 The status of artificial intelligence – whether it could ever become ‘general’, or is destined to remain a ‘narrow’ mathematical manifestation as in the case of current machine learning systems – is a debated issue that goes beyond the scope of the present work. For a non-technical overview, see Broussard 2018.

      2  2 For an overview of AI systems and deep learning methods, see Kelleher 2019. For an introduction to music recommenders, see Celma 2010. Historical accounts and non-technical explanations of machine learning systems can be found in Pasquinelli 2017; Broussard 2018; Sumpter 2018; Natale and Ballatore 2020; and Mackenzie 2015.

      3  3 Chen writes: ‘in 1983, CPU speed or frequency was 25 MHz. Assuming 1,000 CPU cycles correspond to making a decision on a Go board, it would have taken 31 minutes to make a decision about 361 × 360 × 359 = 46,655,640 possible moves back then. Today, a single CPU’s frequency is above 4 GHz, […] making it 160 times faster on the same algorithm – in this case, taking less than 12 seconds to make that same decision’ (2016: 6).

      4  4 Since 2015, special issues on the critical study of algorithms have appeared in recognized social science journals such as the European Journal of Cultural Studies, Information, Communication & Society, New Media & Society, Technology & Human Values, Theory, Culture & Society, and more.

      5  5 The concept of socialization has been at the root of sociological theory since its very beginnings, as a way to explain mechanisms of social reproduction. As Guhin, Calarco and Miller-Idriss (2020) note, the term has become increasingly contested in post-war American sociology. In fact, Parsons’ functionalist notion of socialization has been criticized for being a ‘downloading’ model of culture which downplays the centrality of social interaction. The concept was subsequently employed in different terms by authors such as Giddens, Foucault, Berger and Luckmann, Luhmann, and Bourdieu, and continues to bear theoretical relevance today.

      6  6 For an account of the genesis of the concept of habitus and of its pre-Bourdieusian uses by authors such as Elias, Mauss, Merleau-Ponty and Panofsky, see Sterne 2003 and Lizardo 2004.

      7  7 Deus in machina is a wordplay based on deus ex machina, or ‘god from the machine’ – originally indicating the mechanical systems used to bring onstage actors playing gods in the ancient Greek theatre. I must thank Mauro Barisione, who came up with it during our early email exchanges about this book project.