University of Minnesota Press, book, ISBN 978-1517916527, English, 232 pages, 2025, USA

This book has been eagerly anticipated, as Goriunova has delivered several public talks and keynotes on the topic, sharing aspects of her research and, in a sense, testing them with informed audiences. Even if the reading may resonate with those events (if you attended them), it remains fresh and has achieved a solid coherence in her newly systematised and expanded formulation. Ideal Subjects addresses an urgent topic: the creation and use of abstract subjects by the digital industry and platforms, in relation to physical subjects. Goriunova skilfully speculates on the complex systems of sampling and associating data across time and space – data that originates from individuals and is then abstracted and associated with others into these abstractions. These same abstractions, to which individuals unknowingly belong, subsequently influence social and economic infrastructures, profoundly challenging the historically established notion of subjectivity. Following a lengthy introduction that critically discusses and positions the very nature of the subject, the three subsequent chapters examine the creation of such digital subjects and how computational models, including machine learning, have produced a genealogy of statistically validated ‘ideal subjects’, while representations of the body – our ultimate material reference – are also abstracted associations by LLMs. In one of many timely quotations, Goriunova cites Stengers: “We need to use our abstractions differently”, which emerges as a significant breakthrough to counter these processes, once there is a deeper shared understanding of the scenario we are navigating – something this book helps to establish remarkably well.