Conference Agenda

INTRODUCTION

The digital landscape’s rapid evolution has necessitated a worldwide shift in K-9 curricula, moving from basic ICT skills to fundamental Computer Science (CS) principles. This includes the infrastructure and underlying principles of the internet, such as what the internet is, how devices connect, how data travels, and where online content is stored.

However, a significant global shortage of specialists has forced many teachers without formal CS backgrounds—referred to here as “new CS teachers”—to deliver these concept-focused lessons (e.g., Eurydice, 2022). While children’s knowledge of the internet has been documented as fragmented and experience-based (e.g., Brom et al., 2023; Brom et al., 2025), little is known about the prior conceptions held by these educators. Understanding these teachers’ conceptions is a critical first step in designing effective upskilling programs. For example, a number of lower secondary children believe that information travels through the internet in a single, direct path between sender and receiver, or through one intermediate point only (e.g., Brom et al., 2023), complicating understanding of e-safety rules. If the new CS teachers’ conceptions remain similar to those of these children—particularly regarding the internet “invisible” infrastructure and processes—then upskilling needs to focus on specific conceptual bottlenecks rather than assuming that teachers’ everyday experience with digital technology automatically yields a robust, conceptual understanding. Conversely, if teachers are clearly ahead of pupil trajectories, then teacher education can prioritize pedagogy rather than CS content.

RESEARCH QUESTIONS

The study addresses the following research questions:

• RQ1: What are the conceptions of new K-9 CS teachers regarding fundamental internet principles, including the internet’s structure, access mechanisms, and data storage?
• RQ2: How do these teacher conceptions compare to those held by children in Grades 4, 6, and 8?

THEORY

This study utilizes a Vygotskian knowledge framework (e.g., Edwards et al., 2018) to interpret how individuals build understanding of complex systems. In this model, children initially develop “everyday conceptions” based on direct interaction (e.g., “videos play in an app”), which are later enriched by “scientific conceptions” that explain underlying mechanisms (e.g., video storage on servers). The latter process typically happens in schools. The research also draws on conceptual change theories, such as diSessa’s “knowledge-in-pieces” model—describing knowledge as a patchy mosaic of isolated fragments (diSessa, 2018)—and “knowledge-as-theory” frameworks (Vosniadou, 2013), where experts can be viewed as possessing coherent mental models.

By mapping the developmental trajectories of internet comprehension from childhood to adulthood, this research contributes to both computing education and developmental psychology.

THE TARGET MODEL OF THE INTERNET

The following passage describes an educational target—the model of the internet considered appropriate for average Grade 8 students (e.g., based on national curricula, such as the Czech one). This is also a minimal requirement for new CS teachers.

The model operationalizes the internet as a distributed, global network characterized by the following key components:

• Infrastructure and nodes: The internet consists of millions of interconnected user devices, servers, and “cross-road” devices such as routers.
• Data storage: Storage is recognized as physical and distributed across millions of dedicated servers owned by various entities (companies, government,…).
• Access mechanisms: Wi-Fi is understood as a separate local network where Wi-Fi routers provide wireless connections to devices while being physically connected to the broader internet typically via cables.
• Network structure: Data transmission is a “multi-hop” process where information travels across multiple network devices and intermediate nodes rather than moving directly between users.
• Transmission media: Physical cables, particularly fiber optics, are recognized as the dominant medium for long-distance and undersea communication.

This paper addresses the growing use of digital classroom management tools in compulsory education and examines their implications for behavioral engagement. Classroom management constitutes a fundamental dimension of effective teaching, particularly in young learner classrooms, where sustained behavioral engagement is a prerequisite for meaningful learning (Evertson & Weinstein, 2006; Emmer & Gerwels, 2006). In English as a Foreign Language (EFL) contexts, teachers increasingly employ digital classroom management tools to structure classroom routines, monitor student behavior, and foster participation. Despite the widespread adoption of platforms such as ClassDojo and Classroomscreen across schools, empirical research examining how these tools shape classroom interaction and behavioral engagement.

The overarching aim of this presentation is to critically examine how and under what conditions digital classroom management tools influence students' behavioral engagement. This paper is guided by the following shared research questions:

RQ1: To what extent does the use of digital classroom management tools result in significant changes in the behavioral engagement of second-grade EFL learners, as measured by a pre-test/post-test design?

RQ2: How do second-grade EFL learners and their teacher perceive and describe behavioral changes in classroom participation following the use of digital classroom management tools?

RQ3: In what ways do the qualitative insights from learners and their teacher help explain the changes observed in behavioral engagement scores post-implementation of digital classroom management tools?

The theoretical framework of the paper draws on student engagement theory (Fredericks et al., 2004), which conceptualizes engagement as a multidimensional construct encompassing behavioral, emotional, and cognitive dimensions. Particular emphasis is placed on behavioral engagement, which is related to students' effort, persistence, participation, and compliance with school structures (Davis et al., 2012). In addition, the paper engages with Self-Determination Theory (Deci & Ryan, 1985), especially in relation to the use of digital reward systems and their potential impact on students' intrinsic motivation, autonomy, and perceived fairness. These frameworks allow contributors to critically examine whether digital classroom management tools function as supportive scaffolds for learning or as mechanisms of external control.

Methodologically, the paper features a range of empirical approaches, which include mixed-methods designs, classroom-based interventions, survey research, focus group interviews, and teacher observations. Employing an explanatory sequential mixed-methods design (Creswell, 2014; Ivankova et al., 2006), it was aimed to capture both measurable changes in engagement and the lived experiences of students and teachers.

The intended purpose of the presentation discussion is threefold. First, it aims to synthesize empirical evidence on digital classroom management tools and student engagement by identifying common patterns as well as contextual divergences across international settings. Second, it seeks to critically interrogate the pedagogical assumptions underlying behavior management, particularly the reliance on gamification, surveillance, and reward-based systems. Third, it was aimed at generating practice-oriented implications for teachers, teacher educators, and policymakers.

The current "poly-crisis" landscape—marked by migration, health inequities, and rapid technological disruption—demands that education research re-examine who produces knowledge and how it is trusted. The range of knowledge-producing actors is widening to include data-driven technologies like Generative AI (GenAI). While GenAI promises epistemic interoperability—the seamless transfer of meaning across linguistic and cultural boundaries—it often operates on a logic of universalism that obscures the situated nature of knowledge. This creates a tension between the efficiency of automation and the sincerity and authenticity required for public trust, particularly in high-stakes domains like multilingual health communication.

This research responds to these changing conditions by investigating the contact zones where university students, acting as citizen adaptors, collaborate with Large Language Models (LLMs) to translate complex medical information. Drawing on Donna Haraway’s (2016) concept of sympoiesis ("making-with") and Karen Barad’s (2007) response-ability (the capacity to respond), we challenge the antagonistic dualisms often present in education and translation research: human vs. machine, professional vs. non-professional, and knowing vs. acting. We argue that in the era of deep fakes and algorithmic bias, knowing is not a static retrieval of information but an active, relational practice of care. The study posits that the interoperability of knowledge across cultures cannot be achieved by AI alone; it requires a sympoietic entanglement where human affect and machine momentum co-compose meaning continuously. By engaging students in the adaptation of neurofibromatosis health materials for underserved linguistic communities (Chinese, Indonesian, Russian), we explore how educational settings can transform into sites of "response-able" action.

The theoretical framework integrates the sociotechnical critique of AI momentum with a feminist new materialist ethics of care. We examine how students navigate the frictions of AI, such as its tendency to hallucinate or revert to English-centric epistemologies, not merely as technical errors, but as pedagogical moments that force a "slowing down." This deceleration allows for the emergence of intimacy with the text and the technology, transforming the student from a passive user into an active curator of public knowledge. This study addresses the vital European and international dimension of the conference by proposing a model for citizen adaptation. As national governments increasingly rely on non-governmental actors for knowledge dissemination, this research demonstrates how higher education can equip students to bridge the gap between scientific authority and vulnerable communities. It asks: How can the academic community ensure that the integration of AI in education serves the broader public good? We propose that the answer lies in cultivating a sympoietic sensitivity through a way of knowing and acting that acknowledges our entanglement with machines while fiercely protecting the human nuance required for trust.

When learners engage with generative AI (AI) in isolation, they risk slipping into mechanical modes of interaction—becoming, as one participant described, “just a machine in front of AI.” This study examines how learners move beyond such mechanical patterns through a year‑long collaborative autoethnography (April 2025–January 2026) involving four graduate students and two faculty members exploring how AI becomes integrated into scholarly practice. Rather than striving to optimize either themselves or the tools, participants negotiated how to adapt AI to their individual ways of thinking. Our findings show that sustaining metacognitive agency in AI‑mediated education depends less on prescribed best practices (Holmes & Tuomi, 2022) than on collaborative learning environments foregrounding shared sense‑making and reflexive dialogue (Clarà et al., 2019; Mackay & Tymon, 2013).

Within this collaborative setting, participants documented their AI use through ChatGPT interaction logs and an online group chat while developing their master’s thesis proposals. Following the course, they produced reflective narratives tracing the evolution of their practices, engaged in dialogic commentary through shared documents, and collaboratively identified similarities and differences across accounts. These iterative cycles transformed individual experiences into shared insights about what it means to engage with AI in scholarly work.

Across participants, AI use shifted in ways that reflected differences in how they preferred to think, write, and interact. Instead of aiming to optimize AI performance or maximize efficiency (Long & Magerko, 2020), participants continually adjusted their prompting styles, interaction rhythms, and structuring practices to make AI responses more usable for their cognitive processes. We describe this personalization process as self‑optimization—the ongoing tuning of AI interactions to support one’s habits of thought (Mezirow, 1978; Lave & Wenger, 1991). Evidence of self‑optimization appeared as participants gravitated toward strategies that felt comfortable, reduced frustration, aligned with their preferred conversational pacing, or provided supportive forms of structure. While all participants shaped their AI use over time, the forms of self‑optimization differed depending on disciplinary backgrounds, epistemological preferences (Hofer, 2000), and tolerances for uncertainty (Dahlqvist & Persson, 2023).

Two analytic dimensions shaped these divergent trajectories: (1) preferred approaches to prompt construction, ranging from carefully structured prompts to intentionally rough or provisional ones; and (2) the direction of change over time, which involved either deepening the dialogic process or reshaping the tool to behave in personally meaningful ways. Across these dimensions, students’ interaction patterns crystallized into three primary approaches.

First, Iterative Input–Output Dialogues characterized students who developed ideas through extended conversational exchange. For these participants, appropriating AI involved purifying the dialogic process—using repeated rounds of interaction to articulate ambiguity, surface emerging ideas, and gradually achieve clarity. Their self‑optimization focused on making the conversation itself productive for thinking.

Second, Structured‑Control Interactions were adopted by students seeking predictability and order in AI outputs. These participants used Markdown formatting, templates, and explicit instructions to shape the system’s behavior. Although such practices may resemble tool optimization, the underlying motive was to create an interaction environment that felt cognitively supportive.

Third, Exploratory Trial‑and‑Error Engagement described students who experimented broadly with prompting strategies, frequently incorporating techniques observed among peers. Their personalization involved discovering what worked through continual experimentation, emphasizing flexibility rather than commitment to a single strategy.

Across all three approaches, students’ AI use increasingly aligned with their disciplinary orientations and comfort with uncertainty, yet each trajectory remained distinct. Together, these patterns suggest that learners appropriate AI not by conforming to prompting prescriptions but by tuning AI interactions to the contours of their own thinking—a form of self‑optimization emerging through sustained engagement, reflection, and collaborative sense‑making (Mezirow, 1978; Urquhart et al., 2025).