Generative AI has rapidly transformed the conditions for teaching, studying, writing and learning in higher education. GenAI makes it possible to produce academically sounding texts without engaging in the reflective processes through which understanding traditionally has been formed (Ostenson & Allred, 2025). This development raises fundamental questions about what counts as knowledge, how learning and teaching are understood, and what role higher education should play.

On the one hand, research suggests that reliance on GenAI may reduce students’ capacity for creative and critical thinking, as cognitive tasks are increasingly offloaded to machines (Pikhart & Al-Obaydi, 2025; Toma & Yánez-Pérez, 2024). On the other hand, studies also indicate that students’ creative and critical thinking can be deepened when they actively engage GenAI in their studies (Mollick & Mollick, 2023; Wang & Fan, 2025). Given the lack of conclusive findings, Wegerif and Casebourne (2025) argue that technologies such as GenAI may, on the one hand, open possibilities for deeper understanding, while on the other hand also generating challenges such as dependency, automation of thought, and loss of agency. They suggest that the implications of GenAI use are closely tied to how it is pedagogically framed and integrated in higher education teaching. In other words, this shifts the focus from technology in itself to teaching practices and pedagogical design.

In this study, GenAI is approached from a dialogic perspective (Wegerif & Casebourne, 2025). Dialogue is here not dependent on all participants being conscious, intentional or morally responsible. Rather, dialogue is understood as a pedagogical and phenomenological space in which meaning emerges through encounters with otherness. GenAI can participate in such processes by introducing alternative formulations, unexpected perspectives or conflicting interpretations that invite critical and reflective thinking. From this perspective, learning is understood as a process in which meaning is created in and through dialogue, rather than as the transmission of knowledge. GenAI is therefore not regarded as a tool for delivering answers, but as an actor within pedagogical interactions that can both open and constrain opportunities for reflection and critical thinking.

The aim of this study is to explore how higher education teaching can be developed by focusing on students’ use, understanding, and meaning-making when engaging with generative AI (GenAI) as a dialogic partner. The research question for the study is: How do students use and understand generative AI as a dialogic partner, and how can these practices inform the development of higher education teaching?

The study is grounded in practice-based action research (Carr & Kemmis, 1986) and builds on my teaching in university courses on educational leadership. These courses provide concrete settings where teaching practices involving GenAI are tried out, reflected upon, and gradually developed over time. Within these courses, GenAI is integrated as a dialogic partner in both teaching and students’ academic work, and teaching practices are iteratively developed as part of the action research process.

This study originates from the first author’s experiences of teaching in higher education at a time when generative AI has become an increasingly present part of students’ academic work and in society more broadly. While generative AI involves significant ethical concerns related to sustainability, environmental impact, and authorship, the study is based on the assumption that higher education cannot meaningfully opt out of engaging with such technologies, as they are already widely used in contemporary society. Rather than avoiding generative AI, the study approaches these concerns as issues to be addressed critically and pedagogically in teaching.

Against this background, the study is grounded in practice-based action research (Carr & Kemmis, 1986). The overall aim is to develop the first author’s teaching by integrating GenAI as a dialogic partner in both instruction and students’ academic work. The first author approaches the project primarily as a teacher examining and developing his own teaching practice on a research-informed basis. The second author contributes expertise in AI, theoretical framing, and qualitative analysis and serves as a critical friend throughout the research process. The action research project is conducted across at least three iterative cycles: the autumn of 2025, the spring of 2026, and the autumn of 2026. Each cycle involves the planning, enactment, observation, and reflection of teaching activities that include GenAI as part of regular coursework. Each cycle builds on the previous one, with tasks and AI use adjusted based on what emerges in practice. The research is situated within a higher education programme in Pedagogical Leadership and Development (PLU). Across the courses, the learning objectives focus on developing an understanding of leadership in educational contexts. In the courses, GenAI is integrated into ordinary teaching activities and used in seminar discussions, written assignments, and a concluding reflection, in which students are asked to consider how GenAI has supported or challenged their work in relation to the course learning objectives. All students were informed about the project before the courses began. Participation was voluntary, informed consent was obtained, and students could withdraw at any time without consequences for teaching, assessment, or grading. All student material was anonymised, and only data from consenting participants were included. The study followed the ethical guidelines of TENK and Åbo Akademi University regarding research ethics, anonymity, and voluntary participation. Different student groups participated across the action research cycles. In the first cycle, approximately 30 students at bachelor’s and master’s levels were involved. The second cycle included approximately 15 bachelor’s students. The data was collected during three courses in pedagogical leadership.

The first cycle was exploratory and focused on examining how GenAI could be used to develop students’ understanding of course content through dialogic interaction. Based on the first cycle, there are some preliminary insights, although analysis is still ongoing. One preliminary finding concerns the conditions for pedagogically productive dialogues with GenAI. While students generally engaged with the AI in an interested manner, many dialogues remained relatively short or superficial, according to the student themselves. In practice, many asked the GenAI a question or two, got a well formulated answer and continued the discussion between students. After discussing this in the course, it appears to me as important to have clear structures on the assignment, provide examples of follow-up questions, and guidance on how to continue the dialogue, both in terms of width and breadth. Students also highlighted the need for caution, as the aesthetic and fluent form of AI-generated text may create an illusion of depth, making such texts appear more convincing and factual than they are. Overall, students found it valuable to work with GenAI as part of the course. In the next action research cycle, the teaching will therefore place greater emphasis on providing background knowledge about GenAI as a dialogic partner, what such dialogue can involve, and how to engage in academically oriented discussions with AI.

Carr, W., & Kemmis, S. (1986). Becoming critical: Education, knowledge and action research. London: Falmer. Finnish National Board on Research Integrity. (2019). The ethical principles of research with human participants and ethical review in the human sciences in Finland. Tenk Publications. Mollick, E. & Mollick, L. (2023) Using AI to Implement Effective Teaching Strategies in Classrooms: Five Strategies, Including Prompts. The Wharton School Research Paper. http://dx.doi.org/10.2139/ssrn.4391243 Ostenson, J. & Allred, J. (2025). Thinking and Writing with AI as a Dialogue Partner. The Utah English Journal: 53(19). https://scholarsarchive.byu.edu/uej/vol53/iss1/19 Pikhart, M. & Al-Obaydi, L. (2025). Reporting the potential risk of using AI in higher Education: Subjective perspectives of educators. Computers in Human Behavior Reports. https://doi.org/10.1016/j.chbr.2025.100693 Toma, R.B., Yánez-Pérez, I. (2024). Effects of ChatGPT use on undergraduate students’ creativity: a threat to creative thinking?. Discov Artif Intell 4(74). https://doi.org/10.1007/s44163-024-00172-x Wang, J. & Fan, W. (2025) The effect of ChatGPT on students’ learning performance, learning perception, and higher-order thinking: insights from a meta-analysis. Humanit Soc Sci Commun 12(621). https://doi.org/10.1057/s41599-025-04787-y Wegerif, R. & Casebourne, I. (2025). A dialogic theoretical foundation for integrating generative AI into pedagogical design. British Journal of Educational Technology, 00, 1–16. https://doi.org/10.1111/bjet.70026