Conference Agenda

There is a growing expectation that artificial intelligence will reduce educational inequalities by lowering barriers of language and academic literacy. Within this discourse, AI is framed as a tool that democratizes learning, particularly for students studying in a second language, by compensating for linguistic disadvantage.

This narrative, however, overlooks a parallel development: the rapid expansion of what counts as digital literacy. Digital literacy extends beyond technical skills or task completion to include the ability to interpret, evaluate and critically use diverse forms of information, including text, media and data. While technological systems evolve rapidly and increasingly simplify tasks, the acquisition of these forms of literacy lags behind. As a result, formal higher education training struggles to keep pace, producing a growing gap between technological availability and students’ capacity to convert digital tools into learning, especially among those for whom academic participation already involves working in a second language.

Research on digital inequality has shown that technologies do not enter neutral systems but interact with existing structures of stratification and uneven distributions of resources (Hargittai, 2008; van Dijk, 2020). From a Bourdieusian perspective, access to a resource does not guarantee the ability to convert it into educational advantage (Bourdieu, 1986). Differences in skills, dispositions and familiarity with digital systems shape who can benefit from technological change, even under conditions of formal access.

Recent policy-oriented research further complicates the democratization narrative. The OECD Digital Education Outlook 2026 cautions that generative AI may improve task performance without leading to deeper learning and may widen inequalities when learners lack the cognitive and pedagogical support needed for meaningful use (OECD, 2026).

Against this backdrop, the proposed paper presents a reflexive qualitative analysis of a participatory research seminar conducted with undergraduate students studying in a language that is not their mother tongue. In this seminar, students worked as partners in designing and implementing a qualitative research project, while AI tools were integrated across all stages of learning, including concept clarification, interview preparation, data analysis and analytical writing. The seminar provided an empirical setting for examining how AI becomes embedded in pedagogical relationships and academic practices.

The objective of the study is not to assess the effectiveness of AI as an instructional tool, but to analyze the social processes that emerge through its everyday use in higher education. The study focuses on the formation of a triadic relationship between learner, instructor and technology, and examines how responsibility, dependence and agency are negotiated within this configuration. Particular attention is given to the accumulation of digital capital, understood as the unequal distribution of skills, confidence and interpretive control that enables some students to use AI effectively as a learning resource while others become dependent on its outputs (Ragnedda & Ruiu, 2020).

The guiding research questions are: How do students and instructors negotiate the use of AI within a participatory learning environment? In what ways does AI integration reshape pedagogical guidance, academic authority and student dependence? Under what conditions do AI supported practices mitigate existing inequalities, and when do they reproduce or intensify them?

By approaching AI integration as a situated social practice rather than a technical intervention, the paper contributes to sociological debates on digital inequality, academic literacy and social sorting in higher education. It argues that as AI becomes normalized within academic work, digital literacy shifts from a compensatory resource to a condition for legitimate participation, producing new forms of dependency that intensify existing educational inequalities.

The emergence of generative artificial intelligence (AI) in education represents a critical juncture in the historical relationship between technology, knowledge, and power. Educational institutions worldwide struggle with integrating AI tools into pedagogical practices. This struggle is reflected in the discourse surrounding this integration, which reveals profound ideological tensions about the nature of learning, the role of the state, and the future of human agency within educational systems (Mauta et al. 2025). Amid this friction, the viewpoints of society frequently become eclipsed by those rooted in technological rationalism. Adopting a Foucauldian critical lens, this study interrogates the ways in which policy documents construct and legitimate specific visions of AI in education. By examining how discourse produces social realities, stabilizes power relations, and directs institutional practice, this research aims to unearth the power dynamics and modes of governmentality embedded within these texts.

Foucault’s conceptualization of power/knowledge and governmentality provides a robust framework for analyzing how policies shape the dominant discourse on AI in education (Foucault, 1977, 1982). Central to Foucault's theory is the inseparability of power and knowledge: those who hold power define what constitutes "legitimate knowledge," which in turn reinforces their authority. In AI policy discourse, this could manifest when documents assert claims such as "AI improves learning outcomes" as self-evident truths. By framing AI as a "necessity" for educational futures, these texts could establish a "regime of truth" wherein educators who question AI integration risk being positioned as "outdated" or "uninformed" rather than as holders of valid pedagogical critiques.

Beyond power/knowledge, Foucault’s concept of governmentality, which refers to the 'conduct of conduct', illuminates how AI policies could shape educational practice through rationalized frameworks rather than overt coercion. Governmentality operates by structuring the field of possible actions, making certain choices appear logical and inevitable. In AI policy contexts, technological integration could be framed as an administrative necessity for managing institutional pressures and state-mandated metrics. Rather than being explicitly mandated, AI tools could become embedded within rationalized pedagogical frameworks, creating systems of self-regulation. Students and teachers align their behavior with the "truths" produced by real-time data dashboards, effectively allowing algorithmic logic to govern educational conduct without direct intervention.

The philosophical tensions at the heart of this study concern the dialectic between agency and algorithmic determination. As AI systems increasingly mediate the educational process from personalized learning pathways to automated assessment and behavioral monitoring (Xu & Ouyang, 2022), questions emerge about the erosion of human autonomy (Prunkl, 2024) and the production of what Foucault termed “docile bodies” (Foucault, 1977). The algorithmic governance of learning threatens to transform education from a space of critical engagement and self-formation into a site of optimization, where students and teachers alike become subjects of datafication, prediction, and control (Williamson et al., 2020). This shift resonates with broader neoliberal transformations in education, where knowledge itself becomes commodified, and learning is reframed as a measurable, marketable competency rather than a process of intellectual and ethical development.

These concerns acquire particular significance within the European context, given the region's distinctive regulatory frameworks and humanistic philosophical traditions. The European Union's General Data Protection Regulation (GDPR) and emerging AI Act exemplify regulatory approaches that attempt to balance technological innovation with fundamental rights protections. This study, therefore, examines how different national and supranational policy documents navigate these tensions, revealing the ideological commitments and power structures that shape AI integration in education. Through comparative analysis of policy texts from the United Kingdom, United States, Türkiye, and UNESCO, this research critically examines how "AI in education" gains meaning through discourse.

This study examines how college faculty members negotiate the rapid penetration of artificial intelligence (AI) technologies into teaching and research environments, as reflected in the discourse of an academic WhatsApp group. The study is grounded in sociological and educational literature that highlights the structural tension between two distinct professional habitus in higher education: the academic and the pedagogical. Drawing on Bourdieu’s theory of practice (Bourdieu, 1988, 1990), the academic habitus reflects dispositions oriented toward preserving symbolic capital, disciplinary authority, and standards of scholarly rigor (Matthies & Torka, 2019). In contrast, the pedagogical habitus (Feldman, 2016) embodies dispositions rooted in instructional practice, learner-centered orientations, and concerns for accessibility, reflection, and developmental support. Although Bourdieu did not explicitly differentiate between these two habitus types, subsequent research has documented the divergent cultural logics that guide academic researchers and teacher educators. Studies in teacher education further emphasize the persistent dilemmas and competing identities experienced by professionals situated between these two worlds of action research and teaching (Labaree, 2000; Cochran-Smith & Lytle, 1999; Shulman, 2005). Together, this literature conceptualizes faculty members as operating within a field structured by dual expectations: protecting academic knowledge and advancing pedagogical practice. The entry of AI technologies into teaching and learning environments intensifies these tensions (Weinreb & Yemini, 2023), making the dual habitus especially salient for understanding how faculty interpret, negotiate, and adapt to technological change. The data was collected from a designated WhatsApp group titled “Wisdom in Intelligence”, established on January 1, 2024, to provide an open space for ongoing dialogue about the impact of AI tools on education and teacher training among faculty members in an education college. The group aimed to enable colleagues to share experiences, content, dilemmas, and innovations related to integrating AI tools into teaching and assessment practices. Based on a qualitative content analysis of dozens of messages, the findings indicate that the WhatsApp group functions as a dynamic field in which an ongoing professional struggle unfolds between the two habitus orientations. Rather than a simplistic division of “pro-AI” versus “anti-AI,” a deeper structural divide emerges: the academic habitus activates a strategy of conservation, seeking to protect the value of traditional academic capital, accuracy, originality, explicit knowledge, and mastery of disciplinary language. The pedagogical habitus, by contrast, activates a strategy of conversion, seeking to leverage new technological capital for educational aims: accessibility for disadvantaged learners, removal of learning barriers, scaffolding, and emotional or process-oriented support. This divergence manifests across five dimensions: epistemology, purpose of action, language, symbolic capital, and professional ethics. Thus, the researcher-lecturer evaluates AI through its factual errors and conceptual inaccuracies. In contrast, the teacher-lecturer evaluates their potential to support learners in “focusing,” “understanding,” or “progressing” even when the model’s output is flawed. The findings suggest that AI technologies are not merely tools that disrupt teaching practices; they disrupt professional identities, value systems, and power relations within the academic field. The academic habitus responds through practices of preserving and reinforcing disciplinary boundaries, truth, capital, and distinctiveness, while the pedagogical habitus responds through adaptation, mediation, and inclusion. A form of managed tension emerges between them: an unresolvable yet structurally embedded component of faculty identity in the age of AI. This study highlights that understanding contemporary challenges in learning and teaching requires more than an analysis of technological tools; it demands attention to the embodied dispositions, modes of action, and professional identities, the habitus that educators bring into the academic field.