Book Reviews

Anthony Elliot

The Culture of AI: Everyday Life and the Digital Revolution

London: Routledge, 2019, 268 pp.

The subtitle of Anthony Elliot's The Culture of AI: Everyday Life and the Digital Revolution, recalls Henry Lefebvre's (1947/2000) Critique of Everyday Life, a text which had a subversive impact among critical theorists of the time by bringing to the fore aspects of French society neglected in orthodox Marxism. The similarity, however, is misleading, since Elliot discusses the present as a mere preparation for a future, instead of presenting the future itself as a stake among social forces seeking to control of our daily lives through the digital “revolution”.

The main argument of the book is that, as a transformation of unprecedented scale and intensity is about to occur, we must be ready to rethink much of our daily lives, starting by acknowledging that, as the books concludes, “we might have, as it were, simply run out of styles of thinking or frameworks for understanding the impact of such changes” (p. 200).

This argument is a tricky one because it connects the problem of knowledge (how to know) and the problem of power (what to do) in a problematic way. The essence of the question is nicely expressed by the character Tancredi in The Leopard when he says, “If we want things to stay as they are, things will have to change” (di Lampedusa, 1958/1960: 31) (A more literal translation of the Italian text would go as follows: “If we want that everything stays the same, everything has to change”). The main merit and limitation of this book consists of the way its author addresses this problem. The book comprises an introduction and six chapters.

In the introduction, Elliot defines AI:

[The term AI is] encompassing any computational system that can sense its relevant context and react intelligently to data […] when certain degrees of self-learning, self-awareness and sentience are realized [… and] referring to any computational system which can sense its environment, think, learn and react in response (and cope with surprises) to such data-sensing [emphasis original].

The obvious problem with this definition is that it takes the metaphoric transfer of intelligence from human intelligence to artificial “intelligence” at face value. In other words, we are left wondering what self-awareness – “to sense” or “react intelligently” to data – actually means when applied to non-organic systems. Because the difference between human and artificial intelligence remains unclear, the relative value and limitations of each remain indistinct, leaving no conceptual escape from the “technological tsunami sweeping the globe” (p. 16). The rest of the book describe some features of this tsunami.

In the first chapter, “The digital universe”, Elliot summarises the main points of his analysis, arguing that in this “universe”, social forms of daily life and “complex digital systems” are interdependent when it comes to issues of “transformation” (p. 26). Elliot grounds his analysis in three traditions: the social theory of Nigel Thrift (p. 43); the study of advanced modernisation and the self (p. 45) in Anthony Giddens, Zygmunt Bauman, and Ulrich Beck; and what Elliot calls the “critical discourse […] pertaining to reinvention, innovation and experimentation”. According to Elliot, the latter “calls our attention to an emerging branch of social ideologies of self-fashioning, in which instantaneity, plurality, plasticity, speed and short-termism grip the imaginations of women and men throughout the digitalized cities of the West who are riding the next wave of innovation” (p. 49). For Elliot, these three traditions, “provide a framework within which it is possible to begin to think critically about the emergence and spread of a culture of artificial intelligence [emphasis original]” (p. 51):

By this I mean the general social process by which everyday life and modern institutions become increasingly influenced and shaped by the digitalized and technical apparatuses of AI.

The second chapter, “The rise of robotics”, is about the social, economic and political effects of automated work in the global economy, interpreted through Marx's theory of technology, the arguments of “sceptics” and “transformationalists” about the “fourth industrial revolution”, the influence of globalisation, and offshoring of automation and the resulting disruptions.

In the third chapter, “Digital life and the self”, Elliot discusses the impact of new technologies on the production of the self “and the daily lives of individuals”, arguing that, rather than producing “an enforced solitude”, “new technology creates both new opportunities and new burdens for the self.” (p. 19.) Adopting a psychoanalytical approach, Elliot argues the case for considering the self as an “information processing system” and “construct[ing] our lives as portable selves, moving across society (online and offline) as if the self is an information processor” (p. 84). The grounds for this suggestion is that, “in this age of smart machines, the key psychoanalytic question is not so much how do we connect but what does it mean for the self when we connect” (p. 84). Elliot dismisses Sherry Turkle's concerns about the new solitude and alienation of the digital age, arguing that digital objects such as social media and “robotic pets” contribute to a “transitional space”, offering people opportunities for “engagement with the wider world rather than a defensive reduction of it” (p. 94). For Elliot, the inability to take advantage of these opportunities should not be blamed on new technologies, but on an individual's life conditions, such as “debilitating emotional imprints from their past or because of the impairment or corrosion of their capacity for processing unthought emotion” (p. 102).

In the fourth chapter, “Digital technologies and social interaction”, attention shifts from the individual to the collective impact of new technologies, and Elliot applies Goff-man's “action framework” to the analysis of the institutional organisation of social interaction, the role of chatbots, the impact of co-presence, mobility, desynchronisation, and individualisation on social relationship and Michael Harri's “eclipse of silence”.

The fifth chapter, “Modern societies, mobility and artificial intelligence”, discusses the impact of automation on mobility, as this notion applies to its civilian and military dimensions, such as the “Google car” and “drones and killer robots”.

In the final chapter, “AI and social futures”, Elliot addresses the future of intimacy, health-care, and democracy in order to argue for the necessity of public policy to develop adequate strategies to handle the complex matrix of risks and opportunities associated with transformations in these domains. As intimacy is not limited to “interpersonal relationships”, but includes also “impersonal objects” and “the cultivation of our connection to technology itself” (p. 162), in healthcare the pace of transformation opens up questions only partially addressed by the debate between “technologists” versus “traditionalists”, and democracy is facing “benefits and burdens” that “question liberal, individualist conceptions of democracy” (p. 184). For Elliot, public policy can address these challenges through three main strategies based on the “digital tooling up of an active and engaged citizenry” (p. 195), the role of government, and the role of the markets. The viability of market solutions, however, requires “systemic corrections” or “structural change” towards “greater transparency and accountability” of the companies involved (p. 197), to avoid the risk that “the transformative potential of complex algorithm” would “reproduce and amplify the biases and other human failings to which some critics argue it is ostensibly resistant” (p. 199).

The culture of AI is a useful book for at least two groups of people: those who are not aware of the transformative potential of AI and related technologies, and those inclined to believe the many promises associated with the corporate marketing of this new “revolution”. Elliot's analysis should open the eyes of these people to the complexities, the uncertainties, and ultimately the serious risks associated with the uncritical embrace of these technologies.

For an audience more experienced with a critical intake of the social glitches of technological development, however, Elliot's analysis contains at least a few limitations. The first, as mentioned earlier, is Elliot's definition of AI and the blurring of the difference between human and artificial, life and digital life, and so on.

The second limitation consists of the narrative representation of technological development as an independent variable. Maybe it was a choice to focus on and discuss social implications. But this choice becomes problematic when we discuss the nature of this development – the forces and ideolgies supporting it, the practices legitimising it, and so forth – because the nature of this develepment itself is put outside the range of critical scrutiny. Resistance is futile and compliant adaptation is the only choice. Elliot accurately describes the extension and intensity of the transformative power of new technologies but neglects to identify the sources of this power: the social forces, the ideological myths, and ultimately the nature of the interests that feed the seemingly irresistible penetration of new technologies in our lives.

These two limitations have at least three implications on the discussion about the “culture of AI”. First, this seems a “culture” in which AI is an independent variable of the equation that generates social change. In dis-cursive terms, this brings about the naturalisation of technological development, that is, the interpretation of technological development as a natural phenomenon – a tsunami – that we can deal with only as far as the effects, but not the causes, are concerned.

Second, social conflict and the role of social structure in shaping technological development – a core tenet in the critical theory of technology – are neglected. The “culture of AI” results from technological development construed as a process independent from social forces and impacts a notion of society purged of social conflict. In a critical perspective, this conflict – and not technology per se – is the real engine of social change.

Third, the concept of “revolution” is construed to describe a radical change of established ways of life, conventions, institutions, and even ways of knowing, but not a radical subversion of relations of power. This is important because this kind of technological “revolution” ends up not challenging, but strengthening, existing relations of power.

Elliot's book is useful reading, however, because it reflects an approach or a culture of AI that is influential among corporate and political elites. Students with critical ambitions should familiarise themselves with it if they want to defend the idea (and its implications!) that technological development and other processes apparently “irresistible” – such as globalisation – are not independent from, but dependent on, social structures, relations of power, and ultimately the competition for control over the distribution of values in society.