Conference Agenda

Since the widespread availability of text-generating systems such as ChatGPT from OpenAI, these technologies have received increased attention in educational settings. Teachers use such socio-technical systems for generation of educational content or to adapt materials to different levels. Text-generating systems are also be integrated into daily classroom practice, for example by employing chatbots as interaction partners for students. These algorithmically controlled and platform-based chatbots present new challenges for both students and teachers.
The EU-AI-Act, as well as the national guidelines, regulate the use of AI-based systems in schools. The Standing Conference of the Ministers of Education and Cultural Affairs (KMK) has emphasized in its strategy papers (KMK 2021) that teachers should integrate and critically reflect the functional mechanisms of algorithmic structures, such as text-generating systems in their teaching. Students should be enabled not only to control these systems through prompts, thereby developing AI-competence (Klar and Schleiss 2024), but also to engage critically and reflectively with AI-based text- and image-generating systems.
Within the European discourse on Education(al) Technology in schools, existing research has examined the algorithms and training data of text- and image-generating systems in terms of their respective biases (Baker and Hawn 2022; Eynon 2023; Skilton and Cardinal 2024). However, there is currently a research gap regarding the investigation of platform-based algorithmic systems in educational settings (Cone 2023). As platform architectures contain different modules, they are build on complex arrays of layered structures which are interoperable. They possess agency as they contribute to the ‘world-making’ of education, taking a co-constitutive role as envisioning and inscribing what functionalities are anticipated in education (Cone 2023). Researchers question how platforms’ technical arrangements and their inscriptions configure students as prototypical users in pedagogical settings (Cone 2023), and how pedagogical practices interplay with platform infrastructures (Hartong & Decuypere 2023).
Platforms and AI-based systems are thus not understood as an improvement of schooling but as an omnipresent programmed infrastructure interwoven in everyday practices. With a post-digital perspective that conceptualizes digital technology, in particular platforms, as integral part of everyday practices (Knox 2019; Weich & Macgilchrist 2023), this contribution addresses the question of which biases are reproduced by text-generating systems and how diversity-sensitive teaching can be designed in post-digital classrooms (Swist et al. 2024).
Using empirical material from a participatory action research project, a chat-based role-play simulation is analyzed. There the platform Fobizz enables teachers to integrate pre-configured or self-created chatbots into classroom activities, supporting interactive and role-play based learning scenarios. Through a structured "mega-prompt," the teachers created a forced migration scenario, in which students guide a group of refugees to a safe country. Therefore the teacher defined the chatbot’s role, task, and processes, while also incorporating external digital architectures such as web search, and Wikipedia. The platform provided guidelines for effective prompting, such as including role definitions, step-by-step instructions, and audience-specific information. Plugins and conversation starters further aimed at guiding teachers to improve students' experience.
Although first works on the diversity-sensitive use of AI have already been published, a differentiated analysis of the integration of text-generating systems into classroom teaching with a focus on their diversity-sensitive use is still lacking. Thus, particular attention is paid to the reproduction of heteronormativity and other biases in generated scenarios. In alignment with power-sensitive approaches towards the concept of diversity, this research contributes to a discussion of teachers’ and the technologies’ position within the classroom.
Finally, the contribution critically discusses who bears the responsibility for the use of socio-technical systems in schools and classrooms.

The imaginary of educational technology as universally deployable, irrespective of context, is a discursively constructed and collectively held idea that is increasingly questioned by education researchers and practitioners (Williamson et al., 2024). Edtech, as an imaginary, frames ways of speaking about and doing education that entwines technologies, humans, and organisations within (mostly) private business models, infrastructural dependencies, and industries extending beyond education (Williamson, 2021). Edtech industries tend to spotlight products/services for learning, teaching, and assessments as off-the-shelf goods, ready to be deployed in a fast, standardised, and cost-efficient manner, but obscure regular needs for (in-house) expertise to make them (inter)operable on the ground (González López & Büchner, 2024; Vanermen et al., 2024). "Schools, universities, educators and students", it has been argued, do "not only adopt, but also adapt, resist or refuse various technologies in context-sensitive ways" (Williamson et al., 2024, p. 335). Consequently, those who care about education and technology benefit from context-sensitive approaches that attend to the educational actors and settings involved, as well as to actors that shape edtech industries.

This study aims to contribute to the growing corpus of critical edtech studies by providing a contextualised understanding of the adaptation imaginary from the Chinese edtech industry. So far, critical edtech studies have predominantly contributed to the understanding of edtech in Western and European countries, including edtech start-ups (Decuypere et al., 2024), trade shows (Huang et al., 2025), and a variety of intermediaries that understand, shape, and govern the industry as a so-called ecosystem (Hartong, 2024). However, critical edtech studies have only begun to show the complexities of edtech industries in Southeast Asia (Knox, 2023; Lee et al., 2025), and English-written cases on Chinese contexts remain scarce (Li & Christophe, 2024). At the outset of the Chinese Government's 15th Five-Year Plan (2026–2030), foregrounding science, innovation, and industrial modernisation, it is timely to examine imaginaries of education and technology. Specifically, this study focuses on the policy-endorsed imaginary of "adapting to local conditions" (因地制宜yīndì zhìyí), an idiom originating in classical texts that has long shaped Chinese governance, agriculture, industry, and education, emphasising contextual fit over universal solutions. Recently highlighted by the State Council, adaptation appears as a key strategy for industry to cultivate production and build an innovative, digital economy (Chinese Government Net, 2025).

The study answers the question of how edtech practitioners imagine adaptations of people (e.g., entrepreneurs), technologies (e.g., chatbots), and organisations (e.g., start-ups) from within China. The discipline of Science and Technology Studies (STS) offers theoretical-methodological sensitivities to technoscientific phenomena, allowing one to demonstrate how relations (re)assemble the conditions, compositions, and consequences of education by empirically tracing their making rather than turning to ready-made theoretical explanations (Gorur et al., 2019). Specifically, this study attends to "sociotechnical imaginaries", the collectively held visions of desirable futures that shape and take shape through the material, semiotic, and moral arrangements of everyday life (Jasanoff, 2015, p. 4). A strand of critical edtech studies has been investigating imaginaries and assemblages regarding technoscience and education (Huang et al., 2025; Parola & Grimaldi, 2025), particularly collective visions of what counts as a desirable educational future, often revolving around AI (Knox, 2023; Sun et al., 2025). For example, a study on China's sociotechnical imaginaries demonstrates how AI-enabled edtech is imagined as a driver of national rejuvenation and educational modernisation, positioning technology as a state-guided force for shaping talent, educational reform, and securing future geopolitical leadership (Knox, 2023). In a time of geopolitical tension, hearing the voices of researchers and practitioners from non-European regions is essential to counter universalising assumptions about education (technology) and to foster a contextualised understanding.

This paper develops a sociologically-informed analysis of the contemporary imperative for teachers to be making use of generative AI (GenAI) technologies as part of their work – e.g. to assist in the production of lesson plans, to write student feedback, draft emails and other routine teaching tasks. As such, the paper addresses a deliberately straightforward research question: How are teachers responding to the offer of using generative AI as part of their work?

Whereas educational research has traditionally approached teacher take-up of new technology as a matter of technology ‘adoption’ – i.e. individual rational behaviour influenced by factors such as teachers’ technological competence, confidence and attitudes - this paper contends that this now has limited explanatory power when it comes to the educational take-up of contemporary GenAI technologies.

In this sense, we look to build on recent sociological literature exploring how the take-up of contemporary platformised and datafied technologies is now shaped by market logics that no longer presume free rational choice on the part of individual ‘users’. Indeed, recent studies of the political economy of educational platforms highlights the monopolistic strategies now being adopted by IT industry actors to ‘lock’ users within proprietary platform ecosystems (Birch and Muniesa, 2020; Komljenovic, 2021). In these ecosystems, an individual’s rational considerations around ease of use, efficiency and usefulness still apply but are mitigated (and sometimes superseded) by heuristics such as brand loyalty, sunk cost and learning curve fallacies. As such, any individual commitment to a tool, app or platform is a function of various ongoing ‘investments’: e.g. cognitive effort to achieve technical proficiency, time invested in becoming algorithmically-rewarded (Shuraida and Titah, 2023). In this sense, the propensity of someone engaging with a particular technology can be seen as a quasi-rational behaviour motivated more by habituation and prior investment than explicit intent (e.g. Williamson et al., 2022).

We contend that the latest wave of GenAI tools represent an intensification of these trends – driven by similar principles of proprietary lock-in and perceived investment. The hypothesis that we develop, therefore, is that the GenAI industry is not pursuing teacher ‘adoption’ of its products as is traditionally understood, but instead seeking to propagate an affectively-intense and totalising form of sociotechnical engagement. In this sense, the paper expands on conceptualisations of ‘attachment’ drawn both from Affect Studies and STS (Science and Technology Studies) (Sellar 2015; Clough and Halley 2008; Bucher 2017; Berlant 2011). These studies frame attachment as a ‘structure of relationality’ (Berlant 2011) that sustains more generalised modes of attachment as a collective existence. In this sense, individuals can grow attached to things - e.g. an ideal or a fantasy - at the same time as other people, meaning that the conflicts and negotiations that subsequently ensue have collective implications.

From this basis, the paper develops a political reading of teachers’ attachment to GenAI as a phenomenon that straddles private and public spheres. For example, it highlights the role of AI vendors in creating generalised ‘atmospheres’ of attachment in which the private bleeds into the public (Decuypere and Perrotta 2026), leading to imagined collective needs that can only be satisfied through wholesale institutional commitment to the idea of GenAI (as evident in discourses for schools to ‘get on the AI train’ and similar). In addition, this framing of collective attachment also highlights how what might appear to a wholly individualised form of decision-making is often tied up with ‘bargaining’ strategies where one’s (non)use of the technology in particular circumstance can sustain feelings of shared sovereignty and agency in lieu of the loss of collective political control in other areas of life (Duschinsky & Wilson 2015).

Accelerating social and technological transformations are reshaping higher education systems worldwide, compelling universities to rethink how learning environments are organised, mediated, and governed. Artificial intelligence (AI), particularly generative AI, has emerged as a central driver of this transformation. Rather than functioning solely as an instructional technology, AI increasingly operates as a mediating element of educational space, influencing how students engage with knowledge, structure learning processes, and negotiate academic norms. In this sense, AI contributes to the reconfiguration of learning environments, academic cultures, and the conditions under which educational practices take place.

This paper conceptualises AI as part of the educational space, understood as a socio-material, cultural, and institutional environment in which learning occurs. Drawing on the conceptual tradition of Innovative Learning Environments (Dumont, Istance & Benavides, 2010), educational space is treated not as a fixed physical setting, but as a dynamic constellation of media, norms, pedagogical arrangements, and forms of agency. From this perspective, AI reshapes educational space by mediating interactions between students, knowledge, and institutional expectations, while simultaneously challenging established boundaries of authorship, responsibility, and assessment.

The study adopts a comparative perspective, focusing on Armenia and Scotland – two small countries with strong educational traditions but markedly different institutional trajectories and regulatory cultures in higher education. Scotland represents a relatively mature and policy-oriented higher education system, where digitalisation and AI governance are increasingly embedded in national and institutional strategies. Armenia, by contrast, is situated within a post-Soviet and post-crisis context, where higher education institutions are undergoing ongoing processes of modernisation, and where AI integration often takes place through locally driven, experimental practices rather than comprehensive regulatory frameworks.

This contrast provides a productive lens for examining how educational spaces respond differently to the same technological transformation. Rather than comparing levels of technological advancement, the study explores how AI is incorporated into learning environments, how students adapt their practices within these spaces, and how institutional cultures shape the meanings and boundaries of AI use.

The study is guided by the following research questions:

• How do students in Armenian and Scottish universities use AI in their everyday academic practices, and how do these practices reflect adaptive learning strategies within different educational spaces?

• How does AI contribute to the reconfiguration of learning environments, academic norms, and student agency in higher education?

• What differences emerge between Armenian and Scottish educational spaces in terms of the alignment between institutional frameworks and classroom-level practices related to AI use?

The primary aim of the study is to analyse AI not as a technological innovation per se, but as a mediating force within educational spaces that reshapes learning environments and academic cultures. By foregrounding student practices, the paper seeks to understand how learners actively co-construct emerging norms of AI use and how these norms are stabilized (or remain fluid) within different institutional contexts.

Theoretically, the study draws on three intersecting strands of literature: (1) research on educational space and learning environments, which emphasises the relational and cultural dimensions of learning; (2) media and mediation theories in education, which conceptualise technologies as shaping, rather than merely supporting, educational practices; and (3) sociological approaches to student agency and adaptation in contexts of institutional change. Together, these perspectives enable an analysis of AI as an element of educational space that both reflects and accelerates broader social transformations in higher education.