Conference Program

This contribution uses the antinomical formula (Bertìn, 1968), the methodological pivot of pedagogical problematicism, to investigate two possible postures regarding the use of generative artificial intelligence devices within the school environment and teaching-learning practices, with the aim to overcome both of them towards a solution both ethically and technologically grounded. On the one hand, we have "reliance" as a posture that does not primarily involve critical and reflective thought, but rather it's focus on automation of the practices and knowledge of both educators and learners when using generative artificial intelligence systems. This pole is well represented by the discipline known as learning analytics (Siemens & Long, 2011). On the other hand, the pole of "support" highlights how it is possible to use devices with a critical and reflective approach, not fully entrusting aspects of teaching-learning practices to algorithmic outputs, but intervening where necessary, negotiating meanings, power, and knowledge. This second pole is represented, for example, by minority studies in critical education and technology (Selwyn, 2014).

According to its epistemological framework, pedagogical problematicism posits that the polarities analyzed within the antinomical dialogue possess both positive and negative elements subject to observation and analysis. These elements, therefore, must be taken into consideration with the aim of overcoming the two positions in favor of a synthesis that renounces dogmatic and ideological aspects in favor of greater pedagogical proximity between the positions. The point, therefore, is not to value "support" solely at the expense of "reliance", nor viceversa. This debate runs parallel to the polarized and polarizing positions of the excessively naïve techno-enthusiasts and the neo-Luddites, who would aspire to an almost total return to the analog; it is believed, in this case as well, that both positions have, from a pedagogical perspective, elements to be valued that are useful for the reflection set forth here. Finally, a recent and exploratory perspective will be examined which, thanks to its theoretical framework, overcomes the impasse that these two positions often bring with them. The reference is to the theory of “artificial agency” proposed by Luciano Floridi (Floridi, 2025). This theory, in its originality, does not banish computers nor eminently (all too) human qualities to the disadvantage of synthetic data but, on the contrary, enriches the discussion and represents in nuce a desirable future.

The integration of generative AI in higher education demands a fundamental shift in focus: from what AI produces to how students interact with it. Without deliberate metacognitive scaffolding, conversational AI risks reinforcing cognitive delegation rather than developing critical judgment. This exploratory mixed-methods study examines whether a structured e-tivity embedding the rAIse framework (Reflect-Ask-Interpret-Source-Evolve) as a procedural scaffold can transform AI conversation into a formative and self-regulatory practice aligned with Assessment as Learning principles (Earl, 2003). Two research questions guided the study: RQ1 — To what extent do metacognitive regulation strategies emerge, particularly in self-assessment and revision phases? RQ2 — What interaction profiles emerge from conversational logs, and how do these associate with the degree of alignment between perceived and enacted competence, as revealed by peer feedback?

17 university students (online degree programme in Psychological Sciences, UnitelmaSapienza) participated in an e-tivity titled "Learning to learn with AI," conducted across three synchronous webinars between October and November 2025. In the first webinar, students were introduced to rAIse through live practice and submitted a baseline evaluative prompt formulated without any structured approach, serving as the analytical reference point for measuring subsequent changes in prompt quality and interaction strategy. In the second webinar, students applied rAIse to co-construct an assessment rubric with AI and use it to evaluate their own work through motivated, iterative revision — directly reformulating their baseline prompt using rAIse phases and introducing explicit criteria, meta-prompting, and regulation practices. In the third webinar, each student analyzed an anonymized peer's conversational log using a rAIse-derived coding grid, providing written feedback on prompt quality, cognitive function attributed to AI, and regulation strategies observed; collective debriefing enabled students to confront the gap between perceived and enacted competence. The dataset comprises 68 chat logs, rAIse pre/post questionnaires (four AI literacy dimensions, Likert 1–5), and structured peer feedback forms.

Logs were analyzed using ordinal indicators of prompt quality (0–3), cognitive function attributed to AI (0–3, from substitutive to critical-regulative), transformative iteration level (0–3), and regulation markers including meta-prompting, motivated output critique, and validation requests. Perceived-enacted alignment was estimated by triangulating pre/post questionnaire delta scores with peer feedback evaluations.

Addressing RQ1, metacognitive regulation emerged in three recurring strategies: regulation for clarity (contesting vague AI-generated criteria), regulation for control (meta-prompting, inverting the standard dialogic flow), and regulation for refinement (negotiating content through motivated rejection of AI proposals). Addressing RQ2, five interaction profiles emerged along a delegating-to-regulative continuum: Reflective Learner (29%), Strategic Architect (18%), Task-Oriented (18%), Pragmatic (23%), and Surface Learner (12%). Surface profiles systematically overestimated their competence while reflective profiles showed close alignment between questionnaire delta scores and log-observed strategies. The baseline-to-rAIse comparison further confirmed that structured scaffolding produces qualitative shifts in prompt construction and interaction governance.

These findings yield three design implications: explicit criteria must be embedded structurally in tasks; verification and triangulation must be required and assessed; peer analysis of conversational logs creates a metacognitive mirror unavailable through individual reflection alone. Epistemic responsibility remains with the learner — and the design must structurally demand it.

This paper presents an empirical case study from an Italian lower secondary school exploring the integration of artificial intelligence (AI) within a civic education pathway grounded in Article 54 of the Italian Constitution, which emphasizes duty, responsibility, and service to the Republic. The study investigates how AI can be incorporated into classroom practice while preserving pedagogical judgment and fostering critical thinking, student agency, and democratic learning. The project developed through structured and progressive phases. Students first engaged in reflective civic inquiry through creative and analytical activities, including acrostics centred on constitutional values (duty, honour, loyalty, discipline, sacrifice, oath), autobiographical writing, and the collective composition of a poem. The pathway also included an encounter with the author of a book dedicated to the journalist Antonio Russo, whose professional commitment provided a framework for discussing ethical responsibility and public service. Additionally, students conducted and examined an interview with a witness connected to the figure of magistrate Emilio Alessandrini, introducing authentic testimonial material into the learning process. AI was introduced gradually and framed as an object of reflection rather than as an authoritative source. Within a sociocultural perspective informed by Vygotsky, AI functioned as a mediating tool whose educational value depended on teacher guidance and critical use. In one phase, a collectively written poem was transformed into a song through an AI-based application while maintaining full student ownership of meaning and content. In another phase, AI-supported transcription and summarisation tools were employed during the guided analysis of the recorded interview. Students were encouraged to distinguish between information, interpretation, and personal elaboration, developing awareness of how AI systems reorganise discourse. The case addresses the two dimensions highlighted in this panel. Regarding the teacher’s role, AI integration was based on deliberate instructional design and explicit ethical framing. Decisions about its use were made transparent, ensuring that evaluative and interpretative authority remained human and pedagogically grounded. Regarding student empowerment, the pathway promoted critical AI literacy through dialogical practices inspired by Bakhtin and metacognitive self-evaluation grounded in Flavell’s framework. Structured written reflections following the author encounter supported students in examining both content and learning processes. Recently awarded first prize in a national civic education competition focused on constitutional values, the project demonstrates that AI can contribute to democratic education when embedded in reflective pedagogy and intentional mediation. The findings suggest that critical AI literacy emerges through guided dialogue, ethical positioning, and metacognitive practice rather than through technical proficiency alone. By positioning AI as a dialogical partner requiring critical engagement, this study contributes to discussions on balancing technological innovation with democratic responsibility in educational contexts.

As generative AI becomes embedded in educational practice, a critical question emerges: does it cultivate or erode the metacognitive capacities that underpin democratic, agentive learning? The dominant model of human-AI interaction — in which students ask and AI answers — mirrors cognitive offloading rather than pedagogical dialogue, externalising problem-framing and bypassing the productive struggle through which deep learning occurs. This contribution presents empirical evidence from a quasi-experimental study examining whether a theoretically grounded alternative can foster metacognitive awareness among pre-service teachers, rather than merely replacing their cognitive effort.

The study deployed MAIEUTIC — a conversational AI explicitly architected around Vygotsky's Zone of Proximal Development and Zimmerman's Self-Regulated Learning framework — in an undergraduate Educational Technology course at an Italian university (N=188). Grounded in Bakhtinian dialogism, MAIEUTIC treats the learner not as a passive recipient of information but as an active voice in a genuine epistemic exchange. Rather than answering student questions, it inverts the typical human-AI dialogue: through Socratic questioning and adaptive scaffolding organised across a tri-phasic structure (planning → producing → revising), the system elicits student elaboration at each stage. Crucially, the system incorporates anti-offloading mechanisms — refusing to generate complete answers and requiring elaboration before offering further support — operationalising the idea that the AI's role is to sustain thinking, not to substitute it. Across 25 analysed sessions, MAIEUTIC averaged 6.5 questions per session against students' 1.6, yielding a 4:1 system-to-student ratio — what we term the Interaction Flip. Metacognitive awareness was measured via the MAI-19 (Metacognitive Awareness Inventory, 19 items, 5-point Likert scale) in a matched pre-post design (N=144: MAIEUTIC n=71, ChatGPT control n=73), with participant-generated anonymous codes ensuring privacy while enabling longitudinal tracking.

Results indicate markedly divergent trajectories. MAIEUTIC participants showed substantial growth in self-reported metacognitive awareness (ΔM = +0.625), while the ChatGPT control group showed a slight decline (ΔM = −0.068). A mixed ANOVA confirmed a large Group × Time interaction (F(1,142) = 146.31, p < .001, η²p = .507), which remained robust after adjusting for baseline non-equivalence (ANCOVA, η²p = .404). Nearly half of MAIEUTIC participants (49.3%) met the Reliable Change Index threshold for clinically meaningful improvement; none in the control group did.

These findings suggest that AI dialogue structure — specifically, who asks the questions — is not a neutral design choice but a deeply political and pedagogical one. In educational spaces where democratic deliberation should flourish, designing AI as an interlocutor that questions rather than answers may be a concrete strategy for preserving learner agency, supporting teacher professional identity, and resisting the automation of judgment. The study contributes empirical grounding to theoretical calls for critical AI literacy, and we conclude with implications for educators, AI developers, and policy frameworks that currently treat all AI tools as equivalent.

A major paradigm shift in educational research has rethought the definition of creativity: no longer seen as a fixed individual trait, it is now understood as a complex process stimulable through targeted educational interventions (Runco, 2004). However, opportunities to develop creativity are not equally distributed. Students attending schools in geographically marginalised areas often experience reduced access to cultural institutions, extracurricular activities, and advanced digital resources that can nurture creative development (Mangione, 2024). In these settings, structured educational programs aimed at fostering creativity can play a particularly important compensatory role. While different activities aiming to foster creative practices have been proposed in educational scenarios (Clapham, 2003) , their effectiveness is often hindered by limitations such as conformity and production blocking (Zhou & Luo, 2012) or the prevalence of persistent creativity myths and test-focused environments (Plucker et al., 2011). Current research therefore focuses on finding new methods for efficiently empowering students’ creativity. In this scenario, Generative Artificial Intelligence (AI) systems, when guided by a clear pedagogical direction, have the potential to act as cognitive mediators and facilitators of the creative process, while preserving the centrality of the human individual. This paper presents a study exploring an innovative approach to AI integration for creativity fostering in schools. While much of the literature investigates the impact of AI on the creative Person or Process (Rhodes, 1961), this study shifts the focus to the Product and the Press, the sociocultural context framing the evaluation. Specifically, the research addresses the following question: does the creative artifact produced following the intervention, evaluated by a team of domain experts, reach high levels of quality in terms of creativity?

To address this question, the study adopts a quasi-experimental mixed-methods design. The three-week intervention involves students from marginalized areas of Umbria, Italy, tasked with designing a textual storyboard. To support them, the study utilizes a digital platform featuring a Generative AI-enhanced multi-agent system (MAS), designed as a brainstorming chat between the human participant and artificial agents. The pedagogical framework is grounded in Edward De Bono’s Six Thinking Hats model (De Bono, 1985). The brainstorming sessions feature distinct phases where both the student and agents wear the same hat, promoting parallel thinking and minimizing cognitive conflict. As recent literature highlights, MAS function as communities that enable collaboration among multiple agents to address complex tasks exceeding the capabilities of any single entity, effectively generating diverse perspectives that enrich the collaborative dialogue (Lo Presti et al., 2025).

To evaluate the quality of the final storyboards, a subsample of anonymous artifacts is submitted to independent domain experts using Amabile’s (1982) Consensual Assessment Technique (CAT). This allows for a robust measure of Product creativity grounded in expert consensus, which represents the Press. Finally, implementing this intervention in geographically marginalized contexts highlights the potential of pedagogically structured AI to act as a democratic equalizer, providing equitable access to high-level cognitive scaffolding and fostering critical, agentic learning environments.

The integration of Artificial Intelligence (AI) into educational systems raises significant concerns regarding the automation of pedagogical judgement and its potential impact on educational professionals’ agency. This contribution explores the use of language-model-based multi-agent systems (LLM-MAS) (Cheng, Y., et al., 2024) as cognitive mediators supporting collective ideation and the construction of school visions, avoiding passive delegation to technology. Grounded in a sociocultural framework, AI is conceptualised as a mediating artefact capable of expanding human cognition operating within the Zone of Proximal Development (ZPD), providing adaptive support for tasks that would otherwise be difficult to manage independently (Vygotsky, 1978). In this view, AI does not replace professional judgement but can strengthen agency as the capacity to intervene intentionally in educational processes and act as co-authors of institutional change (Bandura, 2001).

The study presents preliminary findings from a project conducted in three small schools located in central Italy, complex contexts where challenges take locally specific forms and require highly contextualised innovation strategies. Teachers, school leaders and key stakeholders participated in imagination workshops supported by an LLM-MAS designed to complement Design Thinking practices (Brown, 2009; Zampolini et al., 2025a). Unlike generalist language models, the system employs multiple intelligent agents organised with distinct roles (Zampolini et al., 2025b) — Challenger, Jester, Gardener and Conceptualiser — enabling proactive interaction and supporting reflective ideation. Equipped with contextual memory, the system functions not as a mere assistant but as a “cognitive teammate”, participating in problem-setting processes while maintaining human control.

Functional differentiation among agents allows simulation of dialogical and collaborative dynamics, fostering negotiation of meanings and co-construction of knowledge. Such dynamics support professional creativity as a systemic competence emerging from collective problem reformulation and shared solution generation rather than an individual trait (Harris, 2018; Sawyer, 2012).

Analysis of the imagination workshops indicates that participants developed increasingly articulated and context-sensitive school visions through iterative cycles of problem reformulation and solution generation. Interaction with LLM-MAS appeared to facilitate dialogue, the exploration of alternative scenarios and the externalisation of tacit assumptions, contributing to the transformation of imagination from an individual activity into a structured collective process. Rather than replacing human thinking, the system functioned as a cognitive and social scaffold that supported reflective deliberation within the workshops (Belland, 2014). By providing a low-risk environment for experimentation, it also helped reduce barriers to creative engagement, such as fear of error and resistance to ambiguity (Zampolini et. al, 2025a).

Overall, interaction with LLM-MAS supported deeper and more reflective ideation processes, strengthening professional agency and the development of locally grounded solutions. Rather than replacing human thinking, AI functioned as a cognitive and social scaffold that stimulated dialogue, made design alternatives visible, and transformed imagination from an individual capacity into a structured collective practice (Belland, 2014). By providing a safe, low-risk environment for experimentation, multi-agent systems also reduced psychological barriers to creativity, such as fear of error and resistance to ambiguity (Zampolini et al., 2025a).