Conference Agenda

Generative Artificial Intelligence (GenAI) is increasingly shaping educational practice by supporting personalized learning, assisting instructional planning, and automating routine tasks. Alongside these opportunities, GenAI introduces pedagogical, ethical, and organizational challenges, including issues of academic integrity, the reliability and bias of generated outputs, data privacy, and transparency in decision-making (Altinay et al., 2024; Eden et al., 2024). Given that GenAI can influence core teaching activities, its integration requires careful consideration of both classroom practice and institutional policy.

Teachers, as the primary agents of change within classrooms, play a decisive role in determining the success of GenAI integration. Their perceptions, acceptance, and preparedness directly influence the extent to which GenAI tools are meaningfully integrated into classroom practice (Eden et al., 2024; Zhai, 2024). In practice, teachers must decide when GenAI can support learning, how to validate outputs, and how to manage student use while maintaining fairness and accountability. These demands make it important to identify the factors that influence teachers’ acceptance and intention to use GenAI, rather than focusing on adoption as a purely technical choice.

The Unified Theory of Acceptance and Use of Technology (UTAUT) provide a well-established framework for examining such factors (Venkatesh et al., 2003). UTAUT proposes that technology adoption is shaped by four main determinants: performance expectancy, effort expectancy, social influence, and facilitating conditions. Applied to GenAI, performance expectancy concerns whether teachers believe GenAI improves teaching effectiveness, instructional quality, or efficiency. Effort expectancy refers to perceived ease of learning and using GenAI tools. Social influence reflects perceived expectations and support from colleagues, school leadership, and professional communities. Facilitating conditions include access to infrastructure, clear guidelines, and professional development that enable teachers to implement GenAI effectively (Xue et al., 2024; Zhai, 2024).

While UTAUT has been widely used to study teachers’ adoption of educational technologies (Benicio et al., 2024; Pérez, 2024), empirical research addressing GenAI is still emerging. Most of the existing studies focus on digital tools in general, whereas GenAI presents distinct features—generative outputs, uncertainty in responses, and implications for authorship and assessment—that may alter the relative importance of UTAUT constructs. In addition, schools vary in their level of guidance, training provision, and access to approved tools, which may create substantial differences in teachers’ experiences and intentions. These conditions highlight the need for targeted evidence on how teachers evaluate GenAI and what predicts their willingness to integrate it into instruction.

To address this gap, the present study aims to examine teachers’ perceptions and practices regarding the use of GenAI tools in education. Specifically, the study aims to analyze these perceptions through the lens of the UTAUT, and to identify the key factors that predict teachers’ behavioral intention to integrate GenAI tools into their instructional practices. This goal raised the following research questions:

1. How do teachers perceive and use GenAI tools in their educational practice?

2. How do teachers utilize GenAI tools in education, as reflected in the five dimensions of the Unified Theory of Acceptance and Use of Technology (UTAUT)

3. What are the key predictors of teachers’ behavioral intention to use GenAI tools in teaching, as defined by the UTAUT model?

Three years passed since the irruption of Generative AI (GenAI) and its impact on all aspects of our society, especially in education. Since the early stages of its appearance, experts have envisioned that knowledge, communication and information will change forever, labelling this phenomenon as the 4^th industrial revolution. Regarding education, teachers incorporated some changes to their pedagogical practices, either because they choose to take advantage of the new techniques or possibilities that GenAI offers (f.i. incorporating learning activities with AI), or because they felt forced due to students’ use of GenAI (f.i. going back to exams to prevent students from using AI in assessment).

Some studies have been carried out about the phenomenon, most of them analyzing local or institutional experiences. In these studies, researchers found that across higher education, GenAI is expanding unevenly: some university teachers recognize GenAI’s potential but use it only occasionally, citing limited training, support, and clear policies (Wang et al., 2025). However, these studies are focused on one part of the multidimensional phenomenon that GenAI is. Similarly as stated in the TPACK (Technological Pedagogical Content Knowledge) model (Mishra & Koehler, 2006), GenAI in education is affected by the knowledge, uses and stances that teachers have regarding it and to understand the phenomenon and to being able to explain why the teachers are doing what they are doing, a multidimensional study is needed. In this line, Spain funded a research project across the country to analyze the policies and practices of GenAI in universities (PID2023-149069OA-I00). This is a necessary foundation for continuing the work, as a lack of evidence on the present condition of the higher education system hinders the ability to formulate strategies and plan next steps, including those related to training, policy, human resources, and infrastructure.

Despite its massive potential, GenAI is a technological tool and, as it happened with previous technologies, it needs to be studied under the same paradigms. Previous studies about the use of technology in higher education have shown that, both the perception of usefulness of technology and its use, is different according to the academic discipline of faculty members (Mercader & Gairín, 2020). But, what about GenAI? Is the discipline related to the teachers’ knowledge, use and stance? Research so far suggests that GenAI is being used unevenly by faculty across different disciplines: engineering teachers tend to use GenAI more than health, and both more than humanities faculty (Mahmoud, 2025). Some of these reported uses include intelligent tutoring systems, automated assessment, virtual patients/simulators, diagnostic reasoning tutors and personalized feedback for health faculty; and writing feedback, language support, chatbots as conversation partners and inclusive support for diverse learners for social sciences and humanities teachers (Almasri, 2024; Mahmoud, 2025; Velázquez-García, 2025; Wang, 2025). Even though some studies suggest uneven use of GenAI among disciplines, evidence exploring actual potential differences according to the academic disciplines are scarce. The aim of this paper is to analyze the differences of knowledge, use and stances regarding GenAI in university settings depending on the academic discipline of the teachers. To do so, the research questions are:

- What knowledge do university teachers for each academic discipline have on GenAI?

- What are the differences between academic discipline regarding the knowledge of university teachers on GenAI?

- How do university teachers use GenAI for pedagogical purposes in each academic discipline?

- What are the differences in the university teachers’ use of GenAI for pedagogical purposes depending on the academic discipline?

- What are the stances of university teachers regarding GenAI in each academic discipline?

- What are the differences between university teachers’ stances depending on academic discipline?

AI in education is classified as high‑risk under the EU Artificial Intelligence Act (2024). This classification places significant responsibility on schools and teacher education institutions to provide transparent, ethical, and pedagogically grounded AI use. Norway is a relevant case because it is set on becoming the world's most digital country by 2030 (The Norwegian Government, 2025).

This paper explores how pre‑service teachers in Norway navigate the emergence of generative artificial intelligence (AI) in their studies and early teaching practice. While generative AI is rapidly transforming educational landscapes across Europe, teacher education institutions vary widely in readiness, strategic direction, and pedagogical integration. This creates an urgent need to understand how future teachers perceive, adopt, resist, and make sense of AI tools—both as learners and as professionals expected to guide pupils in responsible and effective use.

Our study focuses on three research questions:

1. How and why do pre‑service teachers use generative AI during their studies?
2. How do they perceive their teacher educators’ attitudes and approaches toward AI?
3. How do they expect to use or avoid AI during school practicum?

In the study, we draw primarily on the dual role of the pre-service teacher (See Dolin, Ellebæk & Daugbjerg, 2022). The role in the use of AI changes drastically from teacher to student. For students, the purpose is to use it as a learning tool. For teachers, it is a tool to help them do their job. In our theory section, we will discuss how these two roles may interplay in students' relationship with AI.

For these reasons, we believe that the teacher students may have different approaches and thoughts toward AI than students who are not studying education. This is furthermore interesting in that it might reflect how teachers in the future will (or not) adapt these tools in their teaching.

In recent years, educators, policymakers, families, and society at large have become increasingly cognisant of the importance of creating inclusive learning environments within schools (Florian, 2015). A strong body of international research highlights that inclusive classrooms can foster belonging, enhance engagement, and promote richer learning opportunities by valuing the diverse knowledge, identities, and experiences that students bring (United Nations Educational, Scientific and Cultural Organization (UNESCO), 2020). Developing the mindsets and pedagogical tools needed for inclusive teaching starts at the level of Initial Teacher Education (ITE), where structured preparation can support new teachers in adopting inclusive approaches from the outset of their careers (Ainscow, 2020). However, evidence indicates that many pre‑service teachers (PSTs) struggle to translate their conceptual understanding of inclusion into teaching practice and thus, ITE programmes often encounter difficulties in structuring practicum experiences that meaningfully support the development of inclusive pedagogy as core teaching practice (Rowan et al., 2021).
Virtual Reality (VR) has emerged as one solution to assist in addressing these challenges. For instance, Perinpasingam et al.’s (2023) study of Malaysian PSTs’ participation in VR‑based role‑playing micro‑teaching activities demonstrated that immersive simulation provided an engaging and innovative space for them to practise interacting with diverse learners in a virtual classroom setting, while Karadayı and Evin‑Gencel (2024) report that VR‑supported experiential learning can help PSTs tailor instruction to individual learner differences and thereby strengthen the foundations of inclusive education. Approaches to the provision of VR classroom simulations range from those where an instructor or facilitator directly controls the virtual avatars in what they say and how they act (Finn et al., 2020), to fully automated VR classroom simulations where PSTs engage directly with pre-programmed scenarios and avatars (Huang et al., 2023). Both approaches offer different affordances and constraints; for instance, instructor controlled approaches offer greater flexibility and personalisation to each PST and for the teaching simulation that unfolds, but can be more labour and time-intensive in provision, while automated VR approaches offer greater potential for scaled deployment and more independent usage, but by their pre-programmed nature can be more limited in responsiveness and pedagogical flexibility of the experience.
Recently, Artificial intelligence-enabled Virtual Reality classrooms have emerged as a new avenue of investigation with the potential to bridge the afore-mentioned constraints and affordances of earlier VR classrooms. These virtual classrooms combine immersive VR simulation with artificial intelligence (AI) to create more responsive, adaptive, and unscripted learning environments that go beyond the capabilities of earlier, non‑AI VR systems. For those VR classrooms that relied on pre‑programmed avatars and fixed scenario pathways, for instance, AI‑driven VR now integrates natural‑language processing, real‑time behavioural modelling, and generative pedagogical agents, enabling virtual students to respond dynamically to PSTs’ questions, instructions, and teaching strategies (Docter et al., 2024). These affordances also decrease the need for instructor-controlled VR interventions to provide variety and variation in each individual teaching simulation, thus offering potential for more independent usage by PSTs and scaled deployment across larger cohorts of student-teachers. These systems can provide automated assessment and feedback, allowing PSTs to rehearse and develop complex skills, as King et al. (2025) illustrate in their investigation of functional communication skills development.
Overall, early investigations of AI‑enabled VR classrooms indicate that they can offer increased realism, adaptability, and scalability. In this regard, they hold much potential to provide richer opportunities for PSTs to develop and practice inclusive teaching than has been possible in earlier VR iterations.