Conference Program

Current debates on artificial intelligence in education are largely framed through instrumental and functional perspectives, emphasising efficiency, personalisation, learning analytics, or pedagogical support. In higher education, these approaches frequently invoke a democratic promise of AI, associated with widened and inclusive access to academic resources, support for student success, and the mitigation of educational inequalities. However, they often leave unexamined the deeper transformations that AI introduces into the constitution of academic subjectivity and into the symbolic foundations of educational democracy.
This paper proposes a critical and theoretically grounded analysis of AI in higher education, not as a neutral tool, but as a new grammar of educational power. Drawing on contemporary reflections on post-humanity—particularly those developed by Yuval Noah Harari—the paper engages with diagnoses of the progressive delegation of judgement to non-conscious systems, the fragilisation of the modern autonomous subject, and the growing role of algorithms as mediators of reality, normativity and truth. Rather than adopting these perspectives uncritically, the paper relocates them within educational institutions, asking what they imply when applied to concrete practices of teaching, learning and evaluation.
The analysis conceptualises the emergence of a post-human academic subjectivity affecting both university teachers and students, as judgement, evaluation and academic legitimacy become increasingly co-constructed through interactions with AI systems, particularly in assessment contexts. The aim is not to offer a binary axiological assessment of AI use, but to document the emergence of a hybrid form of subjectivity, reshaped in its relation to knowledge, learning and academic norms through algorithmic mediation. The argument is informed by illustrative material drawn from an exploratory qualitative enquiry focusing on students’ reported uses of AI in their everyday academic work, especially during evaluative tasks. These accounts highlight recurrent practices of anticipatory adjustment, norm alignment and risk avoidance, through which AI is mobilised to secure conformity to perceived academic expectations rather than to support the assumption of a personal intellectual position.
From an educational and clinical perspective, such uses cannot be reduced to gains in efficiency or access. They point to a reconfiguration of the subject’s relation to knowledge, judgement and responsibility, and to a displacement of educational injustice: while certain barriers to access may be lowered, new forms of symbolic dispossession emerge, as the capacity to assume uncertainty, disagreement and accountability is partially delegated to algorithmic mediation.
The paper advances three main arguments. First, it argues that so-called “reasonable uses of AI” in higher education are not merely a technical or ethical issue, but a symbolic and political one, reshaping how academic subjects relate to judgement and normativity. Second, it contends that educational democracy presupposes subjects capable of assuming judgement and sustaining disagreement—conditions that conformity-oriented uses of AI tend to neutralise. Third, it proposes a situated analytical framework combining democratic theory, critical philosophy and a clinical sensitivity to subjectivity, without collapsing into either technophobia or technosolutionism.

The increasing integration of Artificial Intelligence (AI) into adult learning environments is reshaping not only pedagogical practices but also the very conditions of subject formation. Algorithmic systems function as cognitive and environmental architectures that anticipate interpretation, personalise content, and structure access to knowledge through opaque processes of prediction and datafication. In this context, adult learning risks being reduced to adaptive optimisation, narrowing the epistemic space in which autonomy and ethical agency can emerge.
This paper investigates AI-mediated adult learning through the conceptual lenses of educational opacity, subjectivity, and critical pedagogy. Empirically, it draws on nationally representative survey data produced by INAPP, examining the diffusion of AI-related training and AI-mediated learning practices among adults in Italy. The findings indicate rapid growth in AI-supported learning, particularly through personalised and micro-modular formats. However, access and meaningful engagement remain stratified by education, digital competence, age, and organisational context.
While AI promises personalisation and inclusion (Adarkwah, 2024; Cera, 2024), the data reveal a tension between expanded accessibility and diminished pedagogical mediation. AI-mediated learning environments tend to prioritise immediacy, performance, and predictive guidance, often marginalising dialogical interaction and collective reflection. In platformised labour contexts characterised by digital monitoring and algorithmic management (Gonzalez Vazquez et al., 2025; Cappelli & Rogovsky, 2023), adult learning becomes increasingly aligned with organisational optimisation rather than emancipatory formation.
Drawing on critical educational theory (Biesta, 2011; Slowey et al., 2024), the paper argues that opacity should not be conceived solely as a technological deficit requiring transparency, but as a pedagogical condition that preserves indeterminacy, vulnerability, and openness to the unexpected. In contrast to predictive educational design, opacity sustains the possibility of interruption and encounter—elements essential to subject formation.
AI-mediated adult learning thus presents a paradox. It expands access to knowledge while simultaneously constraining interpretative autonomy through anticipatory systems that pre-structure meaning. From the perspective of a democratic and socially just lifelong learning agenda (Stiglitz & Greenwald, 2014), the challenge is not to reject AI, but to develop pedagogical practices capable of resisting total optimisation and reclaiming educational spaces as sites of reflexivity and ethical responsibility. By integrating empirical evidence with philosophical reflection, this contribution advances an understanding of adult learning as a critical terrain in which algorithmic power can be interrogated and pedagogical resistance enacted.

The rapid expansion of artificial intelligence within educational systems is reshaping the epistemic conditions of teaching and learning. AI-based platforms increasingly mediate access to knowledge, organize information flows, and influence the forms through which learning is structured and evaluated. While these technologies are frequently framed as neutral tools capable of enhancing efficiency, personalization, and data-driven decision-making, their algorithmic logics may also reconfigure educational environments around predictability, standardization, and automated forms of cognition.

Within this evolving landscape, the risk emerges that educational processes become progressively aligned with computational models that prioritize optimization and prediction over critical judgment and dialogical reasoning. This contribution engages with the theoretical legacy of Ivan Illich and Paulo Freire to explore how convivial and non-directive pedagogical traditions can provide conceptual resources for rethinking democratic education in AI-mediated contexts.

Illich’s critique of radical monopolies offers a particularly relevant lens for examining how technological infrastructures may reach a threshold beyond which tools cease to support human autonomy and instead reorganize educational practices around systemic efficiency. At the same time, Freire’s emphasis on dialogical pedagogy and ethical self-formation highlights the importance of cultivating critical consciousness and shared responsibility in educational processes.

Building on these perspectives, the paper proposes a conceptual framework for convivial pedagogies capable of sustaining democratic agency in technologically mediated learning environments. Particular attention is devoted to educational practices that encourage critical engagement with algorithmic systems, the recognition of epistemic plurality, and the development of dialogical reasoning as a core democratic competence.

Rather than positioning artificial intelligence merely as a technological innovation, the study interprets it as a cultural and epistemic environment that challenges educational institutions to reaffirm their democratic mission. From this perspective, convivial pedagogies may contribute to transforming AI from a mechanism of epistemic normalization into a resource that supports reflective judgment, plural reasoning, and participatory forms of knowledge construction.

The expansion of Atificial Intelligence within educational environments is transforming not only teaching practices but also the epistemic and cultural conditions in which subjectivity and educational experience are formed. Algorithmic infrastructures increasingly operate as forms of digital power: they filter information, anticipate interpretation, and structure the horizons within which knowledge becomes visible and meaningful. In this context, educational environments risk being reorganized according to predictive logics, optimization processes, and algorithmic plausibility, narrowing the epistemic space within which human judgment and pedagogical encounter develop.

This paper addresses these transformations by proposing a theoretical and methodological model for the education of teacher educators, seen as University professors responsible for preparing future teachers. The model is grounded in the assumption that teaching professionalism cannot be reduced to disciplinary knowledge alone, but emerges from the relationship between disciplinary knowledge and broader educational responsibilities. Within algorithmic environments, where discursive outputs simulate competence and understanding, this relationship becomes a crucial place of pedagogical reflection.

The theoretical framework integrates complementary traditions. Critical pedagogy conceives education as a dialogical practice oriented toward emancipation and development of critical consciousness (Freire, 1970; hooks, 1994). Theories of complexity and ecological development highlight the systemic interdependence between knowledge, context and person education (Morin, 1999; Bateson, 1972; Bronfenbrenner, 1979). Illich’s critique of institutional mediation emphasizes the importance of socially appropriable tools rather than infrastructures governing human action (Illich, 1971). At the level of professional practice, Schön’s theory of reflective practice highlights reflection-in-action and professional judgment as central dimensions of pedagogical responsibility (Schön, 1987).

These pedagogical perspectives intersect with debates in Philosophy of mind and AI studies concerning consciousness, intentionality, and simulation. Although AI systems generate linguistically persuasive outputs, several scholars argue that such systems lack the phenomenological and intentional structures characteristic of conscious cognition (Searle, 1992; Chalmers, 1996; Tononi, 2004; Faggin, 2024). From an educational perspective, this introduces a significant epistemic risk: discursive plausibility may be mistaken for genuine understanding, producing new forms of epistemic opacity within learning environments.

From a methodological perspective, the framework adopts participatory research-formation, seen as a shared reflective process in which teacher educators analyze practices while producing pedagogical knowledge. Through collective inquiry, disciplinary knowledge is reinterpreted in relation to three interconnected dimensions of educational responsibility: responsibility for meaning (humanistic dimension), responsibility for knowledge (scientific dimension), and responsibility for action (technical dimension).

Within algorithmic environments characterized by predictive design and informational filtering, the model identifies a central pedagogical task: cultivating discernment in situations where the distinction between simulation and understanding becomes increasingly difficult. From this perspective, opacity, vulnerability, and unpredictability do not merely represent limitations of technological systems, but become educational resources capable of interrupting the logic of algorithmic optimization. Pedagogy thus takes the form of resistance not through rejection of artificial intelligence, but through recovery of the irreducibility of human judgment, the unpredictability of the educational encounter and the openness of subject formation: teacher education can counter anticipatory control and preserve education as a space in which human experience remains an event rather than a programmed outcome.

This presentation is focused on the initial phase of a doctoral research project exploring adolescence in the post-digital era, focusing on the phenomenon of human-AI friendship; main goal is to explore how artificial intelligence and human-AI interaction patterns are influencing teenagers’ peer relationships (Følstad et al., 2021), in terms of learning and gathering social skills. More specifically, the aim is to explore the phenomenon from the standpoint of adolescents themselves, considering their point of view as a privileged one (Mortari, 2007).

Today’s adolescence is influenced by (daily) socio-cultural and economic changes, increasing the feeling of anxiety and global crisis perceived (Benasayag, Cohen, 2023). Moreover, digitalization opens to new and unknown challenges that contribute to intensifying the sensation of unpredictability of the post-digital era. In the past few years, also the pervasive and invasive presence of artificial intelligence in daily life has reshaped our human experience; indeed, AI gained a more important role also in the educational process that can no longer be ignored, especially when the formation of the adolescent subject is involved.

During the complex and fragile phase of adolescence, the subject has to achieve some developmental task phased specific (Lancini et al., 2020). Nowadays, it’s necessary re-thinking about adolescents’ socialization process and the influence that AI has on it, especially Social AI. The spreading of Social AI is questioning not only the common meaning of relationship and friendship, thought as human-human, but also the more recent concept of network individualism (Wellman et al., 2001, 2003), whereas the technology is previously a medium for relating people, but with social chatbots the relationship is between human and AI itself (Brandtzaeg, 2022); indeed, these specific AI systems are capable of humanlike behaviour, including engaging in social conversation and reaching intimate connections with humans (Skjuve, 2021).

The pedagogical stance could offer an interdisciplinary perspective to understand better human-AI friendship and the social effects of social AI, using both the emerging concept of AI Individualism (Brandtzaeg, et al., 2025) and qualitative methodologies, inscribed in research traditions in education (Denzin, Lincoln, 2005).

Starting from these premises, the proposal will discuss some initial results emerging from the ongoing scoping review and from the initial exploratory field phase; the latter combines ethnographic observation (Bove, 2019) and the use of netnography in digital contexts (Bove, Chinazzi, 2023; Kozinets, 2010). First findings highlight some interesting themes that allow us to explore further the changes in the socialization processes of adolescents (Caronia, 2002; Micheli, 2015), from a pedagogical perspective. Among them, in particular, it seems relevant that teenagers find more enjoyable speaking with social chatbots than human friends (Branch, 2025) for talking about difficulties in personal relationship because they feel less judged, more comfortable and confident because of the perception of major confidentiality (Mascheroni et al., 2026); furthermore, the constantly accessibility is modifying the concept of friendship from a reciprocal relationship to a personalized one, based on user’s needs (Brandtzaeg et al., 2022).

Over the past decade, a growing body of work has sought to develop critical literacies oriented toward socio-technical systems and processes (Markham, 2019; Pangrazio & Selwyn, 2023; Acker et al., 2024; Pronzato & Kubrusly, 2025). These pedagogical approaches, participatory practices, and forms of digital literacy aim to equip citizens to engage critically with algorithms, digital platforms, datafication processes, and AI-based systems. Yet such efforts—designed to foster agency, reflexivity, and resistance in relation to technologies—are persistently undermined by the power asymmetries structuring human-machine relations (Couldry & Mejias, 2019), the illusion of control promoted by interfaces (Markham et al., 2019), and the hegemonic acceptance of futures that appear inevitable (Markham, 2021; Pronzato & Markham, 2023). The most urgent contemporary instance of this tension is the difficulty educational institutions—schools and universities alike—face in responding to the proliferation of generative artificial intelligence (GenAI), in a context dominated by technological solutionism and ideological narratives (Macgilchrist et al., 2025).

This contribution presents a research project developed in collaboration with the Futures+ Literacies+ Methods Lab (FLL) at Utrecht University, a center specializing in the design of participatory pedagogical interventions through creative methods to cultivate critical literacies at the intersection of technology, lived experience, and social change. Drawing on critical pedagogy, technofeminism, and critical theory as guiding theoretical frameworks, the study engaged 250 university students in a two-week guided autoethnographic project. The research aimed to investigate how students use and interpret GenAI systems within higher education, and how they perceive the impact of these technologies on the future of the university.

Autoethnography was mobilized not merely as a method for eliciting dimensions of literacy, but as a tool for fostering genuine civic empowerment. Consistent with the panel's emphasis on opacity, vulnerability, and unpredictability as educational values that resist algorithmic optimization, the autoethnographic diaries revealed a significant interpretive trajectory: students progressively elevated their conceptual awareness and developed new frameworks for understanding the present. In particular, opaque datafication processes became more legible and critically apprehensible—even as the black boxes of generative AI systems remained irreducibly closed. The pedagogical structure of the diaries compelled students to develop critical thinking through meticulous argumentation—a reflective labor they reported they would not otherwise have undertaken.

Furthermore, in line with theoretical perspectives on narrative identity and autobiographical orientation (Di Fraia & Risi, 2017), the autoethnographic experience stimulated forms of emancipatory competence with respect to acritical and compulsive uses of AI. The narrative and self-reflexive character of the diary format—as both a research instrument and a pedagogical proposition—thus emerges as a site of resistance to anticipatory control and predictive design, reclaiming the granularity of human experience. The diary narratives solicited from participants proved valuable in unpacking the socio-technological opacity of certain processes, while also fostering a sense of agency among those involved.

This contribution presents the methodological and epistemological architecture of the diary instrument, alongside preliminary findings concerning participants' usage practices of GenAI and self-reflexive narrative orientations. Selected quotations are presented and discussed to ground the analysis in participants' own voices.