AI & Society最新文献

This article proposes a new, Camusian approach to analyzing and navigating ethical dilemmas in relation to Artificial Intelligence (AI) and, by extension, to other disruptive technoscience. The article takes as its point of departure the Unified Framework of Five Principles for AI in Society, as advanced by Floridi and Cowls (2021), which offers a comprehensive and cohesive framework of the many abstract values and principles brought up in AI ethics discourse. Using a case-study approach, which focuses on the principle of accountability in AI, we demonstrate that, even following an exhaustive systematization of abstract principles, ethical dilemmas still arise whenever we consider applications of the technology in concrete situations. Furthermore, because of the way technology mediates our ethical judgement, a deeper ethical dilemma arises; we can either judge technologies like AI prematurely, without knowing their impact, or after they have been able to bias our norms and intuitions for ethical deliberation. This article then argues that the vulnerability of our ethical judgement to continuous doubt, which is exposed by AI as a landmark case of disruptive technology, can be addressed by integrating Albert Camus’ philosophy of absurdity, rebellion and dignity. Through Camus, we can contextualize our experience of ethical doubt in relation to AI as existentially absurd, while also navigating normative change more confidently with meta-level principles for ethical deliberation itself. The article concludes that, while ethical dilemmas and doubts will persist, this Camusian approach will make us more responsible in undertaking the continuous adaptation and concretizing of our ethical frameworks.

This article interrogates how users interpret and respond to anthropomorphic versus minimalist chatbot designs in legal and regulatory advisory domains, contexts where ambiguity is costly and charm rarely billable. Anchored in ten in-depth interviews and supported by probabilistic simulations employing Bayesian inference and Monte Carlo simulation, the study reveals that interface preferences are far from stylistic whimsy. Instead, they reflect deep-seated expectations rooted in professional roles and interactional demands. Practitioners in law, HR, and compliance consistently gravitate toward pared-down, non-human designs and value transparency, cognitive economy, and semantic precision. In contrast, those operating in branding, UX, or emotionally expressive roles tend to welcome anthropomorphic agents, associating them with engagement and affective resonance. The findings advocate for adaptive chatbot architectures: systems that modulate their aesthetic and communicative cues in response to domain norms, user expectations, and interactional context.

Once emblematic of human intellectual mastery, chess has become a domain where machines not only surpass human ability but fundamentally reshape the practice’s social dynamics, meanings, and power structures. This paper examines the transformative impact of superhuman AI on professional chess, positioning it as an analytical model case for understanding AI’s broader societal implications. Drawing on a corpus of 271 h of transcribed chess commentary, the analysis traces the shift from symbolic AI to deep learning systems, and the consequent reconfiguration of chess as a social practice. The study explores how this transformation alters the game’s meaning, redistributes authority, reshapes power relations, and creates new social roles through AI’s integration. These insights foreshadow challenges for fields, such as medicine or law, where AI’s ascendancy may similarly redistribute authority, redefine purpose, and reshape agency.

What happens when AI systems begin to shape not only our decisions but our sense of self? This article introduces the concept of the institutionalized self—a psychological formation that emerges through recursive interactions with algorithmic feedback in predictive institutional environments. As artificial intelligence (AI) becomes integrated into systems governing education, employment, and healthcare, individuals are increasingly evaluated and categorized by opaque, anticipatory classification systems. These interactions influence not only external opportunities but also internal self-understanding. Drawing on symbolic interactionism, self-determination theory, and technological mediation, the paper proposes a three-stage model of self-formation: institutional perception, metacognitive response, and self-reconfiguration. The model accounts for how institutional classifications affect narrative identity, autonomy, and perceived efficacy. Testable hypotheses are offered to support empirical research, and ethical implications are explored concerning the design of AI systems that shape personal meaning making. The institutionalized self reframes the role of psychology in an era where identity is no longer solely self-authored, but increasingly co-constructed with predictive institutions.

This paper introduces the concept of an algopticon to theorize surveillance in the age of algorithmic governance and surveillance capitalism. Building on earlier surveillance models— panopticon, synopticon, banopticon, and super-panopticon—the algopticon represents a qualitative transformation in how power operates through data, prediction, and automation. Drawing on Hegel’s concept of sublation (Aufhebung), the paper argues that the algopticon does not simply replace these earlier frameworks but sublates them: it negates, preserves, and elevates their core logics into a new surveillance regime. Visibility becomes invisibility; discipline becomes prediction; and observation becomes algorithmic categorization. By automating control and embedding it into everyday life, the algopticon alters subjectivity, restructures agency, and deepens asymmetries of power. Through comparative analysis, this paper shows how the algopticon consolidates disciplinary, synoptic, exclusionary, and informational logics into a pervasive system of behavioral governance. It concludes by emphasizing the ethical and political stakes of this shift and calls for algorithmic accountability, transparency, and democratic oversight to ensure that emerging technologies serve justice rather than entrench inequality.

E.T.A. Hoffmann’s The Sandman (1816) mixed themes such as fear, eeriness and the darkening of the relationship between humans and automatons. The Sandman shows the possible emergence of human machines or so-called artificial intelligence (AI), which is indistinguishable from humans. The paper also looks at emerging and growing technologies, the ethics of artificial life and the nature of consciousness. The technical ability to create automata, developed at the end of the eighteenth century, fascinated nineteenth century authors such as Hoffmann. In The Sandman, Olimpia embodies the nineteenth century fascination with machines and mechanical beings that can imitate and interact with humans. Olimpia is the epitome of the modern concept of AI, in which machines simulate and control human behavior in various capacities. This paper is about how The Sandman anticipates the human ability to understand the nature of reality and technology. With the advent of mass media and the development of AI, I show how the protagonist Nathaniel complicates the conceptualization of the human subject. By relying on and compromising with technology, Nathaniel loses his humanity and turns himself into a madman. The paper also highlights the text and compares it to contemporary AI narratives. Fear and anxiety, reflections on the loss of humanity in the face of technological advancement and the potential of AI to surpass human intelligence are also discussed.

Enhancing algorithmic transparency is a pivotal issue in the ethics of artificial intelligence. From a phenomenological perspective, algorithmic transparency comprises two dimensions: the openness and visibility of algorithmic systems and the subject’s cognitive engagement with understanding the operational logic and theoretical underpinnings of algorithms. Consequently, algorithmic transparency should be reframed as a cognitive issue, with education serving as the central pathway for enhancing cognitive capacity. Unlike traditional educational paradigms that prioritize knowledge transmission, phenomenological pedagogy emphasizes intuitive experience and cultivation of critical thinking, enabling learners to achieve deeper understanding and foster the construction of knowledge. This approach not only underscores the practical value of phenomenological pedagogy in advancing algorithmic transparency but also highlights that such transparency depends not merely on public disclosure of algorithmic information but on learners’ experiential intuition and reflective analysis. Compared to purely technical solutions, improving transparency through the enhancement of learners’ cognitive comprehension of algorithms offers distinct and transformative advantages.