What Is the “Human” in Human-Centered AI? Restoring the Aesthetic Subject

Abstract
Human-centered artificial intelligence (HCAI) seeks to orient AI toward human values, autonomy, oversight, safety, and meaningful control. Yet the term “human-centered” leaves a prior question unresolved: what conception of the human is being centered? Much HCAI discourse approaches the human as user, decision-maker, rights-holder, stakeholder, or bearer of preferences and values. These conceptions are indispensable but incomplete. This article argues that HCAI must also recognize the human as an aesthetic subject: a being whose perception, attention, imagination, interpretation, and judgment are formed through experience. As recommendation, personalization, generative AI, and predictive systems increasingly organize the conditions under which people encounter and evaluate the world, AI can participate in forming rather than merely serving human preferences. The article develops the concept of aesthetic agency to describe the human capacity to participate meaningfully in the formation of one’s perceptual, imaginative, interpretive, and evaluative relation to the world. It distinguishes choice from judgment, preference from taste, generation from imagination, and optimization from experience. On this basis, it proposes an aesthetically expanded HCAI organized around openness, plurality, reflectivity, and transformability. To center the human is not only to preserve control over AI, but to preserve the capacity to perceive, imagine, judge, and become otherwise.

Keywords
Human-Centered AI; Aesthetic Subject; Aesthetic Agency; Aesthetic Judgment; Human–AI Interaction

Cite this paper
Byungkil Choi, What Is the “Human” in Human-Centered AI? Restoring the Aesthetic Subject , SCIREA Journal of Sociology. Volume 10, Issue 2, April 2026 | PP. 102-143. 10.54647/sociology841594

References

[ 1 ]	Ben Shneiderman, Human-Centered AI (New York: Oxford University Press, 2022), especially the discussion of the two-dimensional HCAI framework; Ben Shneiderman, “Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy,” International Journal of Human–Computer Interaction 36, no. 6 (2020): 495–504, especially 495–498.
[ 2 ]	Immanuel Kant, Critique of the Power of Judgment, ed. Paul Guyer, trans. Paul Guyer and Eric Matthews (Cambridge: Cambridge University Press, 2000), §§1–9; John Dewey, Art as Experience (New York: Minton, Balch & Company, 1934), chs. 2–3. Kant is cited for judgment that is not mechanically derived from determinate concepts; Dewey for experience as temporally and environmentally constituted rather than reducible to preference satisfaction.
[ 3 ]	Don Ihde, Technology and the Lifeworld: From Garden to Earth (Bloomington: Indiana University Press, 1990), especially the chapters on human–technology relations; Peter-Paul Verbeek, What Things Do: Philosophical Reflections on Technology, Agency, and Design, trans. Robert P. Crease (University Park: Pennsylvania State University Press, 2005), especially Part II on technological mediation.
[ 4 ]	Shneiderman, Human-Centered AI, discussion of the two-dimensional HCAI framework; Shneiderman, “Human-Centered Artificial Intelligence,” 495–498.
[ 5 ]	Dewey, Art as Experience, chs. 2–3, especially the discussion of organism–environment interaction and “an experience.”
[ 6 ]	International Organization for Standardization, ISO 9241-210:2019, Ergonomics of Human-System Interaction—Part 210: Human-Centred Design for Interactive Systems (Geneva: ISO, 2019).
[ 7 ]	Joseph A. Konstan and John Riedl, “Recommender Systems: From Algorithms to User Experience,” User Modeling and User-Adapted Interaction 22 (2012): 101–123, especially 101–103 and the discussion of user-experience evaluation. The cited claim—that embedding the algorithm in the user experience affects its value and that evaluation requires measures beyond predictive accuracy—is stated explicitly at the outset of the article.
[ 8 ]	Luciano Floridi et al., “AI4People—An Ethical Framework for a Good AI Society: Opportunities, Risks, Principles, and Recommendations,” Minds and Machines 28 (2018): 689–707, especially 696–700; Virginia Dignum, Responsible Artificial Intelligence: How to Develop and Use AI in a Responsible Way (Cham: Springer, 2019), especially the chapters developing responsibility and human-centered governance. Floridi et al. formulate beneficence, non-maleficence, autonomy, justice, and explicability, with explicability joining intelligibility and accountability.
[ 9 ]	Solon Barocas, Moritz Hardt, and Arvind Narayanan, Fairness and Machine Learning: Limitations and Opportunities (Cambridge, MA: MIT Press, 2023), especially the chapters introducing group fairness criteria and their normative assumptions; Jon Kleinberg, Sendhil Mullainathan, and Manish Raghavan, “Inherent Trade-Offs in the Fair Determination of Risk Scores,” Proceedings of the 8th Innovations in Theoretical Computer Science Conference (ITCS 2017), Article 43, 43:1–43:23, especially the impossibility results.
[ 10 ]	Joshua A. Kroll et al., “Accountable Algorithms,” University of Pennsylvania Law Review 165, no. 3 (2017): 633–705, especially 660–682 on transparency, procedural regularity, and technical mechanisms of accountability; Brent Mittelstadt et al., “The Ethics of Algorithms: Mapping the Debate,” Big Data & Society 3, no. 2 (2016): 1–21, especially the sections “Map of the Ethics of Algorithms,” “Transformative Effects,” and “Traceability.”
[ 11 ]	High-Level Expert Group on Artificial Intelligence, Ethics Guidelines for Trustworthy AI (Brussels: European Commission, 2019), especially the sections on human agency and oversight; Shneiderman, Human-Centered AI, discussion of human control and human responsibility.
[ 12 ]	Immanuel Kant, Critique of the Power of Judgment, ed. Paul Guyer, trans. Paul Guyer and Eric Matthews (Cambridge: Cambridge University Press, 2000), §§1–9, especially the account of judgments of taste as grounded in feeling rather than determinate concepts while nevertheless laying claim to universal assent.
[ 13 ]	Martha C. Nussbaum, Love’s Knowledge: Essays on Philosophy and Literature (New York: Oxford University Press, 1990), especially the essays on literary imagination and moral perception; Iris Murdoch, The Sovereignty of Good (London: Routledge & Kegan Paul, 1970), especially “The Idea of Perfection.”
[ 14 ]	John Dewey, Art as Experience (New York: Minton, Balch & Company, 1934), chs. 2–3, especially the distinction between ordinary experience and “an experience” and the account of organism–environment interaction.
[ 15 ]	Pierre Bourdieu, Distinction: A Social Critique of the Judgement of Taste, trans. Richard Nice (Cambridge, MA: Harvard University Press, 1984), especially Part I on the social conditions of taste; Dewey, Art as Experience, chs. 2–3.
[ 16 ]	Alexander Gottlieb Baumgarten, Aesthetica, 2 vols. (Frankfurt an der Oder: Johann Christian Kleyb, 1750–1758), §1. See also Alexander Gottlieb Baumgarten, Metaphysics: A Critical Translation with Kant’s Elucidations, Selected Notes, and Related Materials, trans. and ed. Courtney D. Fugate and John Hymers (London: Bloomsbury, 2013). Baumgarten famously formulated aesthetics as scientia cognitionis sensitivae, the science of sensible cognition.
[ 17 ]	Immanuel Kant, Critique of the Power of Judgment, ed. Paul Guyer, trans. Paul Guyer and Eric Matthews (Cambridge: Cambridge University Press, 2000), §§1–9, 18–22. For the broader distinction between determining and reflecting judgment, see the First Introduction and the published Introduction, especially §§IV–V.
[ 18 ]	John Dewey, Art as Experience (New York: Minton, Balch & Company, 1934), chs. 2–3, especially the account of “an experience,” doing and undergoing, and the temporal integration of experience.
[ 19 ]	Maurice Merleau-Ponty, Phenomenology of Perception, trans. Donald A. Landes (London: Routledge, 2012), Preface and Part One, especially the analysis of the lived body as the medium of perceptual access to the world.
[ 20 ]	Don Ihde, Technology and the Lifeworld: From Garden to Earth (Bloomington: Indiana University Press, 1990), especially the analysis of human–technology relations; Peter-Paul Verbeek, What Things Do: Philosophical Reflections on Technology, Agency, and Design, trans. Robert P. Crease (University Park: Pennsylvania State University Press, 2005), Part II on technological mediation.
[ 21 ]	Dietmar Jannach, Markus Zanker, Alexander Felfernig, and Gerhard Friedrich, Recommender Systems: An Introduction (Cambridge: Cambridge University Press, 2010), especially the chapters on evaluation and user-oriented recommendation; Joseph A. Konstan and John Riedl, “Recommender Systems: From Algorithms to User Experience,” User Modeling and User-Adapted Interaction 22 (2012): 101–123, especially 101–103 and the discussion of user-experience evaluation.
[ 22 ]	Byung Cheol Lee and Jaeyeon (Jae) Chung, “An Empirical Investigation of the Impact of ChatGPT on Creativity,” Nature Human Behaviour 8 (2024): 1906–1914, especially the Results and Discussion. Across five experiments, the authors report that ChatGPT assistance increased the creativity of ideas relative to no technology or conventional web search, with stronger effects for incremental than radically new ideas.
[ 23 ]	Jack McGuire, David De Cremer, and Tim Van de Cruys, “Establishing the Importance of Co-Creation and Self-Efficacy in Creative Collaboration with Artificial Intelligence,” Scientific Reports 14 (2024): Article 18525, especially the Results and Discussion. The study shows that the creative effects of generative-AI collaboration depend importantly on the organization of human–AI interaction and perceived co-creation.
[ 24 ]	Michelle Vaccaro, Abdullah Almaatouq, and Thomas W. Malone, “When Combinations of Humans and AI Are Useful: A Systematic Review and Meta-Analysis,” Nature Human Behaviour 8 (2024): 2293–2303, especially 2295–2299. The meta-analysis of 106 experimental studies reports negative average synergy for decision tasks but significantly more favorable effects for open-ended creation tasks.
[ 25 ]	Qazi Mohammad Areeb et al., “Filter Bubbles in Recommender Systems: Fact or Fallacy—A Systematic Review,” WIREs Data Mining and Knowledge Discovery 13, no. 6 (2023): e1512, especially the review findings and the sections on mitigation and explainable recommender systems. The review finds evidence of filter-bubble effects while emphasizing multiple contributing biases and the potential of diversity, transparency, explainability, and user control to mitigate them.
[ 26 ]	Immanuel Kant, Critique of the Power of Judgment, ed. Paul Guyer, trans. Paul Guyer and Eric Matthews (Cambridge: Cambridge University Press, 2000), especially the Introduction, §§IV–V, and §§1–9. Kant's distinction between determining and reflecting judgment is important here because reflective judgment confronts particulars for which an adequate universal is not simply given in advance.
[ 27 ]	Hannah Arendt, Lectures on Kant's Political Philosophy, ed. Ronald Beiner (Chicago: University of Chicago Press, 1982), especially 42–45, 70–72. Arendt develops Kantian judgment through the ideas of enlarged mentality and the capacity to consider possible standpoints of others.
[ 28 ]	Raja Parasuraman and Dietrich H. Manzey, “Complacency and Bias in Human Use of Automation: An Attentional Integration,” Human Factors 52, no. 3 (2010): 381–410, especially the sections synthesizing automation complacency, automation bias, and their attentional mechanisms; Mary L. Cummings, “Automation Bias in Intelligent Time Critical Decision Support Systems,” in AIAA 1st Intelligent Systems Technical Conference (Reston, VA: American Institute of Aeronautics and Astronautics, 2004). These studies concern automation rather than generative AI specifically and are used only to support the broader claim that formal human authority does not eliminate overreliance on automated recommendations.
[ 29 ]	Pierre Bourdieu, Distinction: A Social Critique of the Judgement of Taste, trans. Richard Nice (Cambridge, MA: Harvard University Press, 1984), especially Part I on the social conditions and class structuring of taste. Bourdieu is used here as a corrective to any conception of aesthetic autonomy as the expression of a socially unconditioned individual subject.
[ 30 ]	Kent F. Hubert, Kim N. Awa, and Darya L. Zabelina, “The Current State of Artificial Intelligence Generative Language Models Is More Creative Than Humans on Divergent Thinking Tasks,” Scientific Reports 14 (2024): Article 3440, especially the Results and Discussion. The study supports the limited methodological point that aesthetic agency should not be grounded in an assumption of categorical human superiority on conventional divergent-thinking measures.
[ 31 ]	John Dewey, Art as Experience (New York: Minton, Balch & Company, 1934), chs. 2–3, especially the account of “an experience,” doing and undergoing, and temporal development. Dewey is cited here for the claim that aesthetic quality concerns the organization of an experience rather than isolated moments of immediate satisfaction.
[ 32 ]	Ben Shneiderman, Human-Centered AI (New York: Oxford University Press, 2022), discussion of the two-dimensional HCAI framework; Ben Shneiderman, “Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy,” International Journal of Human–Computer Interaction 36, no. 6 (2020): 495–504, especially 495–498. Shneiderman’s framework is relevant because it explicitly rejects a simple trade-off between automation and human control.