Conference Agenda

The conceptual mastery of chemical equilibrium remains one of the most persistent pedagogical challenges in secondary and post-secondary science education. This topic requires students to navigate complex cognitive shifts between the macroscopic world of observable phenomenaand the invisible, sub-microscopic world of particulate interactions (Rajmawati et al., 2022; Chiu et al., 2002). This difficulty is frequently framed within the triplet model of chemical representation, which posits that deep understanding relies on the learner's ability to seamlessly integrate the macro, sub-micro, and symbolic levels of chemistry (Talanquer, 2011). Research consistently indicates that students often struggle to connect these levels, frequently relying on rote memorization of algorithmic procedures without developing a conceptual schema of the underlying molecular dynamics (Ncgobo & Moloi, 2025).

To address these disconnects, educational technology has increasingly turned toward digital visualizations. For decades, interactive simulations have been established as effective tools for visualizing the invisible, allowing learners to manipulate variables and observe the immediate consequences on molecular behavior (Olympiou et al., 2013). Tools such as PhET simulations have become staples in the classroom, credited with reducing cognitive load and helping students build mental models of abstract concepts (Rahmawati et al., 2022). However, a critical limitation remains in how these tools are typically implemented. Educational research suggests that while exploring a simulation promotes inquiry, the act of designing and building a simulation engages higher-order cognitive processes, forcing learners to explicitly define the relationships between variables (Magana & Jong, 2018).

This distinction aligns with the theoretical framework of constructionism, which posits that learning is most effective when learners are actively engaged in creating tangible public artifacts (Papert & Harel, 1991). In the context of science education, computational thinking and coding offer a pathway for this constructionist learning. However, the practical integration of coding into chemistry curricula has been hampered by a significant barrier: the syntactic burden of programming (Fuchs et al., 2024).

The emergence of Generative Artificial Intelligence (GenAI) proposes a paradigm shift to overcome this barrier. Large Language Models (LLMs) have the potential to handle the syntactic complexities of coding (Jiang et al., 2024), thereby freeing students to focus on the semantic logic of the scientific content. In this arrangement, AI transitions from a tool for information retrieval to a co-coding partner. This partnership creates a unique Zone of Proximal Development (Chaiklin, 2003), where the student provides the scientific reasoning and the AI provides the technical execution. This allows non-coding students to become creators of scientific artifacts, potentially deepening their engagement with the triplet of chemical representation by forcing them to describe particulate behavior in precise, algorithmic terms.

Furthermore, the effectiveness of such interventions is often amplified when grounded in relevant contexts. As argued by Rahayu (2019), integrating Socio-Scientific Issues (SSI) into chemistry education is essential for enhancing chemical literacy and transferable skills. The carbonic acid–bicarbonate buffer system in human blood provides an ideal SSI context, linking the abstract concept of equilibrium shifts to the tangible health imperative of pH stability .

Despite the proliferation of GenAI tools, there remains a paucity of empirical research on how students navigate the specific intersection of prompt engineering, chemical logic, and simulation design. While the potential for AI-as-partner is theorized, the specific dynamics of how students correct AI misconceptions and refine their own understanding through this dialogue are not yet fully understood. This study addresses this gap by investigating an instructional intervention where students utilize AI to design interactive simulations of the blood buffer system, moving from passive consumers to active architects of their own learning tools.

Digital media are widely regarded as a key prerequisite for enabling students’ participation in a digitized society, and their sustainable integration into classroom teaching is a central concern of educational policy across Europe (cf. European Commission, 2023). In this context, tablet-based group work has gained prominence as a didactic arrangement expected to support new forms of learning, collaboration, and knowledge production in response to changing learning cultures. While output-oriented research suggests that digital technologies can be associated with gains in competencies or learning outcomes (cf. Chen et al., 2018), considerably less is known about how learning and teaching processes are accomplished in digitally mediated classroom interaction.

From a didactic and classroom-interaction perspective, this constitutes a significant research gap. In particular, the operational level at which teaching and learning unfold through interactional practices has rarely been examined in relation to the material specificities of digital technologies (cf. Cerratto Pargman & Jahnke 2019). Existing research often conceptualizes tablets, apps, or digital platforms as tools or resources supporting learning, rather than as constitutive elements of classroom practice (cf. Kalthoff/Röhl, 2011). As a result, the interactional organization of group work and the role digital materialities play in shaping epistemic activity remain insufficiently understood (cf. Engel & Jörissen 2022, but: Mathieu, 2021).

This paper addresses this gap by adopting an inductive-reconstructive perspective on digitized classroom practice (cf. Proske et al. 2023). It draws on findings from a DFG-funded research project investigating how digital materialities participate in social processes of knowledge production in tablet-based group work. The central objective is to contribute to didactic research by analyzing learning and teaching not primarily as outcomes or instructional designs, but as interactionally organized, sociomaterial practices through which knowledge is collaboratively produced in classrooms, with a particular focus on recurrent demands and challenges in group-based epistemic work.

The paper is guided by the following research questions:

• (1) Which recurrent demands and challenges arise when groups collaboratively accomplish knowledge products within arrangements of teachers, students, and digital materialities?
• (2) How do task characteristics and features of the digital environment shape the interactional practices through which groups deal with these demands and challenges?
• (3) How do digital materialities participate in these practices, and how is this participation made interactionally relevant?

The theoretical framework combines interaction-analytic approaches with sociomaterial perspectives. From an ethnomethodological perspective, classroom learning is understood as a situated accomplishment emerging through participants’ coordinated actions, talk, embodied conduct, and the use of material artefacts (cf. Housley, 2012). Sociomaterial approaches conceptualize digital technology not merely as external tools, but as active components of human-technology arrangements that co-shape educational practices (cf. Fenwick et al., 2011; Nevile, 2014; Valasmo et al. 2022). Bringing these perspectives together, the paper introduces epistemic vision as an analytic heuristic building on Goodwin’s concept of professional vision (Goodwin, 1994), capturing how participants orient to and make relevant epistemic objects under conditions of dynamic, screen-based materiality.

The pedagogical uncertainties created by generative artificial intelligence (GenAI) since 2022 is profoundly challenging didactic practices at an unprecedented speed. In uncertain times, there is an urgent need for teaching innovations to go beyond technical instruction and develop learners’ critical, ethical and epistemic literacies. While student adoption of GenAI is widespread in higher education, institutional responses have often centred on prohibition or narrow compliance guidance rather than pedagogical grounded engagement. Emerging research on critical AI literacy argues that such approaches are insufficient as they fail to develop learners’ capacity to make informed judgements about appropriate use (Bearman et al., 2024). Furthermore, researchers have called for collaborative and reflective pedagogies to be embedded in didactic practices in order to enhance learners’ higher-order thinking (Li et al., 2025).

This study investigates how dialogic, game-based learning can support higher education students in their development of critical GenAI literacy and feedback literacy within a design-based methodology that explicitly embraces pedagogical uncertainty. The conceptual framework integrates four strands. First, dialogic theory positions learning as meaning-making through encounters between multiple voices rather than transmission of fixed static knowledge (Bakhtin, 1981; Linell, 2009). Feedback literacy, therefore, requires reciprocal and exploratory exchanges that can open potentialities for learning when students are invited to question, justify and co-construct understandings with others (Steen-Utheim & Wittek, 2017). Second, ongoing work on student and shared epistemic agency conceptualises agency as distributed across people, tools and institutional systems, rather than as an individual capacity (Damşa et al., 2010). This is especially salient in GenAI contexts where students’ sense of authorship and responsibility is negotiated with non-human actors, including assessment practices (Yang et al., 2024). Third, the study is underpinned by sociomaterial perspectives which foreground how students, teachers, technologies, spaces and policies configure what can be known and achieved in learning environments (Fenwick et al., 2011). The GenAI analogue card game designed for this study, therefore, can be regarded not as a neutral pedagogical artefact but as an actor that shape possibilities for action. Fourth, the study is underpinned by game-based learning theories which highlight how games can create playful, low-risk environments where players experiment, iterate and reflect, particularly when debriefing and guided reflection are built into design (Plass et al., 2015). In other words, games can serve as dialogic spaces where players (learners) discuss, debate and reason their decisions (Hanghøj et al., 2021). To this end, the present study addresses the following research questions:

1. To what extent does participation in a game-based learning session improve students’ critical GenAI literacy and feedback literacy?
2. In what ways does the dialogic game mediate the development of learner agency for ethical and responsible AI use?