Full text
Repugnance to Futility: Predictive AI and the Eclipse of Action Sassan T. Asl Working Paper – Version 1 Date: December 2025 Abstract Recent debates on algorithmic AI often frame public anxiety in terms of fear, risk, or ethical harm. This paper proposes a different diagnosis. Drawing on Hannah Arendt’s political theory, it argues that contemporary reactions to predictive AI are better understood as a form of repugnance toward futility rather than anxiety about danger. Predictive systems increasingly optimize, anticipate, and to some extent automate social processes, yet they do so by suppressing the conditions of action that Arendt associates with freedom, plurality, and natality. While such systems are highly productive in terms of outputs, they tend to foreclose the emergence of genuinely new beginnings, reducing political and social life to the management of probabilities. The paper conceptualizes predictive AI as a form of organized futility: a mode of technological governance that functions efficiently while eclipsing the space of appearance in which action becomes possible. By reframing contemporary AI backlash through an Arendtian lens, the paper contributes to ongoing discussions in political theory and science and technology studies about algorithmic governance, democratic encounter, and the limits of prediction. It also situates the concept of repugnance as an affective response to the erosion of action, rather than as a rejection of technology per se. Introduction The rise of predictive AI systems has transformed how human behavior is anticipated and managed across social and political spheres. Conventional analyses often frame reactions to AI in terms of ethical risk, bias, or fear of automation. This paper contends that another perspective is necessary: the sense of futility engendered by AI’s capacity to optimize without allowing for genuine beginnings. Hannah Arendt’s concepts of action, natality, and the space of appearance provide the
theoretical lens to understand this emotional response. Unlike physical threats or ethical lapses, predictive systems subtly constrain political and social action by foreclosing unpredictability and suppressing plurality. Predictive AI and the Eclipse of Action Predictive infrastructures, such as recommendation algorithms, policing prediction tools, and large-scale deliberation platforms, operate by transforming future possibilities into probabilities. While highly efficient, this transformation reduces the space for political action and social encounter. Helping from Arendt, this paper argues that such systems induce a form of “repugnance to futility”: an affective response to the recognition that the technological optimization of social processes limits the emergence of genuinely new initiatives. This repugnance is not fear of AI, but an ethical-political sentiment concerning the erosion of conditions necessary for plurality and democratic encounter Conclusion By conceptualizing predictive AI as organized futility, this working paper establishes a theoretical bridge between political philosophy and the study of algorithmic infrastructures. It emphasizes the importance of maintaining the space of action and plurality, even in highly automated contexts. The present document represents Version 1 of this working paper and is intended as a foundation for ongoing development and refinement. References ·Arendt, H. (1958). The Human Condition. University of Chicago Press. ·Zuboff, S. (2019). The Age of Surveillance Capitalism. PublicAffairs. ·Amoore, L. (2020). Cloud Ethics: Algorithms and the Attributes of Ourselves and Others. Duke University Press.
·Small, C., et al. (2021). Polis: Scaling Deliberation by Mapping High Dimensional Opinion Spaces.
