Conference Agenda

Students’ and teachers’ competencies are pivotal for effective use of educational technology, such as generative artificial intelligence (GenAI) (Van der Vlies, 2020; Gisbert Cervera & Caena, 2022). However, the concept of AI literacy (Long & Magerko, 2020; Ng et al., 2021) overlooks three GenAI-specific aspects: 1) GenAI is trained to interpret and generate natural language, and thus also reflects social determinants of language (such as social class, ethnicity and gender) (White et al., 2023). Previous research overlooks that competent and conscious prompting of AI tools should account for that sociolinguistic dimension. 2) Another caveat in previous research concerns the importance of the iterative nature of GenAI prompting, as competence does not lie in the initial optimal prompt, but in the refinement of prompts throughout the iterative process (Bozkurt & Sharma, 2023). 3) Most research has focused on competences of humans using GenAI tools, while the aspect of non-human agency to influence social dynamics in GenAI tools should not be overlooked (Latour, 1996; Akrich, 1992). Competent prompting is therefore also a sociotechnical process. Drawing on qualitative data from teachers and students, this paper introduces prompting (l)iteracy as a novel concept for competent human-GenAI interaction. This paper is part of the Chat Equity project of the European Research Council (funded by the European Commission under grant number 101168540).

The research focuses on three questions: 1) Which GenAI tools (Large Language Model)- and which features within that software are used in secondary education in Flanders (Belgium)? 2) What roles do GenAI tools take in education? And 3) What school policies exist in Flemish secondary schools to manage the usage of GenAI in both teachers and students? In all three research questions, special attention is paid to the usage by multilingual learners, and the role schools take to support them through LLMs.

This paper qualitatively conceptualizes prompting (l)iteracy in secondary education drawing on 10 focus groups with 25 teachers and 25 students, and 50 individual interviews with teachers and students. During the focus groups, the general usage of LLMs for education is discussed in focus groups with teachers on the one hand and students on the other hand. In these focus groups, people with various levels of familiarity with GenAI tools are included. Which AI tools do they use and for which purposes? The focus groups provided us with a comprehensive view of GenAI usage in and for secondary education. During the individual interviews, teachers and students with experience using GenAI for education were selected. Each of the respondents was invited to bring two conversations they had with GenAI, including their initial prompt and the iterative conversation that might have followed. The conversations are elaborated upon in the interviews to make the cognitive process while using GenAI tangible. The data from focus groups and individual interviews was analyzed using NVivo in the light of Social Practice Theory (Reckwitz, 2002; Shove, Pantzar & Watson, 2012).

This paper maps the GenAI tools used for educational purposes by teachers and students in secondary schools. It dives into the roles LLMs play for teachers including the development of learning and evaluative materials, LLMs as a support to provide students with feedback and LLM support for multilingual students in language development and the acquisition of other knowledge and skills in a language that is not their first language. It looks into how GenAI tool features steer interaction and classroom norms. With the data on students, we mapped to what extent LLMs are used as search engines, to reveal answers to assignments or how they can play a role as study buddy.

Digitalization has become a key policy focus across Europe, with EU member states increasingly encouraged to implement and regulate digital practices in schools. According to an Early Childhood Education and Care (ECEC) policy survey in 2022, 50% of participating countries support the use of digital books in ECEC (Liang et al., 2025). While digital literacy materials have been shown to enhance skills in a single language (Fuernes et al., 2021; Hare et al., 2024), they also hold potential for fostering multilingual competence or language awareness, as advocated in plurilingual approaches promoted by EU member states. Furthermore, they may support ECEC practitioners, who are sometimes reported to feel uncertain about which multilingual approaches to adopt (Platzgummer & Thoma, 2023) and who lack appropriate materials. In Sweden, for example, the limited availability of children’s books in languages other than Swedish has led some schools to introduce a picture book app to promote language and literacy development in multiple languages (Nilsen, 2024). In Luxembourg – a small country with Luxembourgish, German and French as its official languages – preschool teachers are required to teach Luxembourgish, offer three weekly activities in French, and develop children’s language awareness. To address the lack of suitable ECEC materials in multiple languages, teams in Luxembourg, Italy, Slovenia and Switzerland produced digital stories in many languages along with activities designed to develop language skills, language awareness, intercultural awareness, and early literacy skills. Their project partners in Germany and Greece created an online platform through which all materials are freely accessible.

This qualitative study examines the ways in which six preschool teachers in Luxembourg implemented digital multilingual storytelling activities as well as children’s participation during these activities. Preschool education is compulsory for pupils aged 4-6 and only one third of the enrolled children speak Luxembourgish at home. To accommodate children’s diverse linguistic needs, teachers have been shown to use languages flexibly and engage in translanguaging practices (Kirsch, 2020) throughout the day and during analogue literacy practices.

Previous empirical studies have demonstrated that both analogue and digital shared reading can enhance vocabulary and grammatical development in one language (Furenes et al. 2021; Pillinger & Vardy, 2022; Reich et al., 2016). Combined shared reading in school and home languages can support the development of skills in two languages, provided interactions are of high quality (Grøver et al. 2022; Larson et al. 2020). Studies on analogue and digital reading, mainly interventions, have also highlighted the role of the adult and their support in one or, more rarely, two languages (Grøver et al. 2022; Liang et al., 2025; Pico & Woods, 2023; Takacs et al. 2014). Findings from natural settings show that emergent bilinguals and teachers routinely translanguage during analogue or digital literacy activities (Nilsen, 2024; Sembiante et al., 2023), which facilitates communication and language development and supports identity-construction.Overall, however, studies on digital reading have paid comparatively little attention to the role of the adult, interactions in multiple languages, and outcomes beyond language skills. In response to these gaps, the present study investigates in what ways and in which languages preschool teachers in Luxembourg engage children in digital storytelling activities and in what ways children participate.

The findings of the current study deepen our understanding of the tensions practitioners experience when engaging emergent multilinguals in digital storytelling activities in languages that are not school languages and unfamiliar to them. They are particularly timely given the reported increase in the use of digital media to promote language and literacy development in schools, alongside teachers’ expressed feelings of insecurity in this area (OECD, 2023).

Academic reading is a cornerstone of successful participation in higher education, serving not only as a means of accessing disciplinary knowledge but also as a central academic literacy practice through which students construct meaning, develop disciplinary identities, and engage in academic discourse communities (Afdal et al., 2023; Hyland, 2023; Pecorari & Shaw, 2022). Despite its importance, research across European higher education contexts indicates that academic reading is frequently assumed rather than explicitly taught, leaving many students underprepared for the cognitive, linguistic, and metacognitive demands of university study (Carillo, 2021; Haukås et al., 2023; Murray, 2022). This issue is particularly pronounced in English-medium instruction (EMI) programmes, where students are expected to process complex academic texts in an additional language while adapting to Western academic conventions (Klimovich et al., 2023; Li & Primecz, 2023).

Within Central and Eastern Europe, Hungary has emerged as a key destination for international students, especially in medical and science-related EMI programmes (Li & Primecz, 2023; Tempus Public Foundation, 2023). Students entering these programmes often come from diverse educational and linguistic backgrounds and may lack prior exposure to the implicit expectations of academic reading in European higher education. Addressing this gap aligns with broader European priorities concerning equity, inclusion, and the development of transferable academic skills in internationalised university contexts.

From a theoretical perspective, academic reading is conceptualised in this study as a cognitively and metacognitively regulated process. Cognitive models of reading comprehension emphasise that understanding results from the active construction and integration of textual information with prior knowledge and task goals (Kintsch, 1998). Building on this view, research has highlighted the crucial role of metacognitive regulation—namely planning, monitoring, and evaluating comprehension—in successful academic reading (Flavell, 1979; Baker, 2008). Skilled readers do not process texts uniformly; rather, they allocate attention selectively, monitor comprehension, and strategically reread text segments in response to task demands (Pressley & Afflerbach, 1995).

Metacognitive regulation is particularly salient in EMI contexts, where readers must manage additional linguistic demands alongside disciplinary comprehension (Csíkos, 2022; Haukås et al., 2023). Explicit metacognitive reading instruction aims to support this process by making regulatory strategies visible and teachable, enabling learners to engage in more goal-directed and autonomous reading practices (Pecorari & Shaw, 2022). Although empirical studies consistently report positive effects of such instruction, much of the evidence relies on self-report instruments, which capture learners’ retrospective perceptions rather than real-time reading behaviour (Veenman & van Cleef, 2019).

To address this limitation, the present study adopts eye-tracking as a process-oriented method for examining reading behaviour. Eye-tracking is based on the assumption that eye movements reflect underlying cognitive activity and attentional allocation during reading (Rayner, 2009; Hyönä, 2010). Measures such as fixation duration, regressions, and rereading patterns provide behavioural indicators of comprehension difficulty and metacognitive monitoring (Godfroid, 2020; Holmqvist et al., 2011). Because eye-tracking captures reading as it unfolds, it enables direct examination of whether metacognitive instruction leads to observable changes in online reading behaviour (Conklin & Pellicer-Sánchez, 2016).

Grounded in this theoretical framework, the study conceptualises academic reading competence as a developmental construct shaped through instruction and practice (Steklács, 2014).

Situated within a Hungarian EMI context, the study contributes to European academic literacy research by examining how international students develop metacognitively regulated reading behaviour in multilingual higher education settings.

Accordingly, the study addresses the following research question:

How does explicit metacognitive reading instruction influence the academic reading behaviour of international university students in Hungary, as measured through eye-tracking?