Conference Program

The rapid integration of artificial intelligence into educational contexts often outpaces shared, critical pedagogical reflection. Dominant approaches to AI literacy tend to prioritize technical and operational skills; however, when everyday AI use is not supported by intentional and reflective practices, it risks fostering forms of cognitive delegation (cognitive offloading). This, in turn, can weaken students’ metacognitive reflection, sense of responsibility, and agency within democratic learning environments.

This contribution draws on a situated educational experience in upper secondary education (grades 10–12) and is based on participant observation of a cycle of workshops in which AI is not introduced as a tool for providing answers or optimizing learning outcomes, but as an object and interlocutor of critical dialogue. Through structured activities comparing human-produced and AI-generated texts, collective discussion, and guided reflection, students are invited to interrogate not only the correctness or plausibility of responses, but also the cognitive and emotional processes activated in their interactions with AI systems.

Across the activities, AI use emerges as responding not exclusively to cognitive needs, but also as a source of immediate reassurance and emotional containment. Especially during adolescence—when identity formation, performance anxiety, and emotional uncertainty tend to intensify—AI often becomes a first interlocutor to whom questions, doubts, and decisions are delegated. While this immediacy may reduce uncertainty, it can also limit students’ engagement with doubt, error, and the inherent complexity of autonomous thinking. From this perspective, AI can be understood as a mediating object that, if not made explicit and critically examined, risks shifting from a transitional support to a stable substitute for reflective processes.

The pedagogical approach adopted is grounded in a sociocultural and dialogical perspective, in which AI is framed as a potential mediator of learning rather than a replacement for educational mediation and pedagogical judgment. Within this framework, metacognition plays a central role: students are supported in making their thinking processes explicit, distinguishing between automatic reliance and deliberate choice, and developing reflective responsibility in the use of AI as a support for thinking and decision-making.

The observations collected consistently indicate that dialogical practices of this kind foster active participation, critical awareness, and student agency. Rather than limiting learning to the acquisition of technical competencies, such practices sustain processes of deliberation, dialogue, and positioning. In this light, AI is not framed as a threat to autonomy, but as a pedagogical opportunity to rethink and strengthen dialogical practices oriented toward conscious participation and democratic learning.

This contribution engages with the panel Critical Dialogue with AI in Schools by foregrounding critical dialogue and metacognition as essential safeguards for nurturing agency, reflective responsibility, and democratic participation amid AI’s growing presence in schools.

Over the past years, sociological research has extensively investigated the use of digital devices and its implications, highlighting both problematic and positive aspects (Busch and McCarty, 2021). While identifying the determinants of problematic digital use has proven complex, age consistently emerges as one of the most relevant explanatory factors (Van Deursen et al., 2015).

Building on this literature, the present study focuses on both ascriptive and acquired characteristics to examine which individual-level variables contribute to differentiated profiles of problematic smartphone use, while considering differences in digital skill levels and competencies in the use of digital devices (Gui and Argentin, 2011).

This study investigates the relationship between problematic smartphone use (Giancola and Salmieri, 2020; 2024) and direct exposure to digital risks among individuals aged 17 to 31. It is based on an original survey designed to explore digital habits, perceptions, and forms of vulnerability associated with mobile device use. The analytical framework combines intensity of use - measured as hours spent on the mobile phone - with two composite indices: one capturing problematic use and the other measuring experiential exposure to the negative consequences of digital life. A multivariate approach is adopted to identify differentiated vulnerability profiles.

The core hypothesis is that intensity of use, perceived problematic use, and directly experienced digital risks are interrelated dimensions of digital vulnerability. To disentangle this complex interplay, the analysis follows four steps: (1) the construction of an index of intensity of use, an index of problematic smartphone use, and an index of digital social risk; (2) the examination of these indices across individual characteristics of respondents; (3) the investigation of their interrelationships through bivariate and trivariate analyses; and (4) the estimation of regression models to identify the determinants of problematic smartphone use and to clarify the interactions between use intensity, perceived problems, and experienced risks.

Overall, the study contributes to the literature by moving beyond an individual-pathology perspective on problematic smartphone use and by conceptualizing digital vulnerability as a relational and structural phenomenon emerging from the interaction between behavioural patterns, social positioning, and experiential exposure within contemporary digital environments.

Research to date has largely examined the intensity and type of digital use in relation to well-being and mental health (Twenge & Campbell, 2018; Orben & Przybylski, 2019), while less attention has been devoted to the long-term implications of age at first access for these outcomes (Dempsey et al., 2020). Research addressing the relationship between early digital access and students’ academic and digital competences remains scarce (Respi et al., 2025; Rioseco-Pais et al., 2024; Gui et al., 2020). In this work, we investigate how early exposure to different digital technologies is associated with learning performance and digital competence, and how socio-economic, cultural, and territorial differences shape these dynamics. The study draws on INVALSI data from the Grade 10 (cohort 2024/2025). In addition to standardized test scores in Italian and Mathematics, the dataset includes information on students’ current and past digital use and a newly developed performance-based measure of digital competence aligned with the European DigComp framework, providing a more robust alternative to self-reported indicators commonly used in large-scale surveys (Siddiq et al., 2016; Giganti, 2024). To examine the relationship between age of first access to digital technologies and learning outcomes, we estimate multivariate regression models controlling for students’ socio-economic and cultural background (ESCS), gender, migrant background, school track, and geographical area. Models include both linear and non-linear specifications in order to capture potential threshold effects, and results are adjusted through weighting procedures aimed at improving comparability across groups characterized by different levels of early exposure. Standard errors are clustered at the classroom level to account for shared contextual effects. Findings on the average age of first access to smartphones, messaging apps, and social media are consistent with previous literature. Moreover, students from Southern Italy and those with a migrant background tend to access digital devices at earlier ages, particularly smartphones, messaging apps, and social media platforms. Early access to social media and messaging apps is consistently associated with lower achievement in both Italian and Mathematics. The magnitude of these differences is comparable to, and in some cases larger than, the well-documented gender gaps in Italian and Mathematics. For smartphones, however, the pattern is non-linear: postponing access up to lower secondary school is associated with higher scores, whereas further delay corresponds to a slight decline. A notable exception emerges for early access to personal computers, which is associated with higher academic performance. The relationship with digital competence is more complex: intermediate access to smartphones and messaging apps appears to be associated with the highest levels of competence, while both very early and very late access correspond to lower scores, suggesting a non-linear pattern in which early familiarity does not automatically translate into stronger digital skills. By contrast, access to social media accounts and personal computers follows patterns similar to those observed for academic achievement. Interaction analyses reveal subgroup heterogeneity, particularly by gender, school track and migrant background, although results remain uneven across technologies and outcomes. This study provides evidence to inform educational policies and contribute to reducing digital inequalities.

This study examines the relationship between social media use and adolescent mental health in the European Union drawing on PISA 2022 data. The study moves beyond “screen time” narratives by analyzing intensity of social media use, gender disparities, contextual vulnerabilities, and policy responses under the EU’s regulatory framework.

The first theme addresses the pervasiveness of social media use among 15-year-olds in Europe. With 96% of adolescents engaging in social media daily and 37% spending more than three hours per weekday on passive browsing activities, social media has become a structural component of youth socialization. The study will contextualize these findings within broader digital transformation trends and adolescence as a sensitive developmental phase marked by cognitive, emotional, and social reconfiguration. It will highlight the distinction between moderate and intensive use, noting that excessive engagement, defined as more than three hours per day, is consistently associated with poorer well-being outcomes.

The second theme explores the correlation between intensive social media use and self-reported depression and anxiety. Drawing on regression analyses from four EU Member States (Hungary, Ireland, Slovenia, and Spain), the study shows that adolescents engaging in high-intensity social media use are significantly more likely to report symptoms of depression and anxiety compared to low- or moderate-use peers. Importantly, the analysis controls for socioeconomic background, academic performance, family support, sense of belonging at school, and experiences of bullying, thereby strengthening the robustness of the association. While causality cannot be firmly established, the evidence underscores a meaningful and policy-relevant correlation.

A third central theme concerns gender disparity. Female adolescents display both higher rates of intensive social media use and substantially higher prevalence of depression and anxiety symptoms. However, the marginal increase in mental health risk associated with higher use appears broadly similar across genders, suggesting that time spent online interacts with other structural and psychosocial factors. This section will argue for gender-sensitive approaches that recognize differential vulnerabilities without oversimplifying the phenomenon as exclusively female-driven.

The fourth theme focuses on adolescents’ perceptions of digital regulation. Evidence indicates strong resistance to blanket bans on devices or restrictive filtering measures, while demonstrating support for collaborative rule-setting with teachers. These finding foregrounds youth agency and suggests that participatory governance models may be more effective than prohibition-based approaches. The study will connect these insights to ongoing EU policy initiatives, including the Digital Services Act, which mandates risk assessment and mitigation of systemic harms, and the broader agenda on youth mental health protection in the digital sphere.

Finally, the study outlines future research and policy directions. It calls for multidimensional interventions combining platform accountability, digital literacy education, parental and school engagement, and longitudinal EU-wide monitoring. It also emphasizes the need for improved access to platform data to better understand behavioral mechanisms and algorithmic exposure patterns. By integrating empirical evidence, youth perspectives, and regulatory developments, the proposed study aims to contribute to a balanced, evidence-based framework for safeguarding adolescent mental health while preserving the social and participatory benefits of digital environments.

constitutes a figurational transformation whose implications for democratic knowledge production remain theoretically underspecified. Drawing on Norbert Elias's figurational sociology and the concept of technological re-enchantment developed in relation to Weberian rationalisation theory (Císař Brown, 2025), the paper argues that AI integration is restructuring the interdependence figuration of the university in ways that generate a structural form of epistemic dependence incompatible with the reality-adequate knowledge production Elias associated with civilising processes and functional democratisation.

The argument consists of three pillars. First, it examines how Elias's concept of reality congruence, and the graduated detachment it requires, presupposes a figuration in which the transmission and validation of knowledge remains at least in principle reconstructable by participants. Second, it analyses how AI systems, through their subsymbolic architecture and recursive optimisation logic, produce what is characterised here as a structural explanation deficit: not contingent ignorance remediable through better pedagogy, but a constitutive opacity that renders epistemic submission a rational adaptation rather than an aberration. Third, it considers what this means for the Eliasian question of functional democratisation: specifically, whether higher education can continue to produce the capacity for autonomous, self-regulating epistemic judgement on which democratic participation depends when the figuration of knowledge production is being reorganised around interfaces rather than understanding.

The paper is careful to avoid technological determinism: the structural conditions described are products of specific developmental choices rather than inherent properties of AI as such. Nevertheless, it argues that these choices have produced a figuration whose civilising and de-civilising implications deserve sustained sociological attention.