Conference Agenda

Generative AI tools like ChatGPT, Claude, and Gemini are rapidly becoming part of how education researchers write, think, and produce knowledge. Most discussions treat these tools as technical assistants: useful but neutral additions to the research process. This study challenges that view. It argues that generative AI should be understood as a knowledge actor, something that actively shapes what counts as authorship, originality, and trustworthy knowledge in education research.

The timing matters. Education research today operates under multiple pressures demands for productivity, visibility, and policy impact, combined with challenges to academic credibility from predatory publishing and misinformation. Generative AI enters this already fragile ecosystem, blurring the lines between human and machine contributions, between originality and recombination, between assistance and substitution.

Current policy responses focus mainly on compliance: disclosure requirements, plagiarism detection, and author responsibility statements (Committee on Publication Ethics, 2023; Elsevier, 2024). Major publishers now require authors to declare AI use but prohibit listing AI as a co-author (SAGE Publishing, 2024; Springer Nature, 2024). These are necessary steps, but they address symptoms rather than deeper questions. What remains underexplored is how AI fundamentally changes the epistemology of education research: how it transforms what we understand as knowledge contribution, intellectual work, and scholarly expertise (Messeri & Crockett, 2024).

This study examines three interconnected questions:

1. How does AI use reconfigure the concept of authorship in education research?
2. How is originality being redefined when AI can generate, combine, and reorganize text?
3. How can academic trust be sustained when the boundaries between human and machine contributions become unclear?

The research uses Actor-Network Theory (Latour, 2005) as its primary framework, treating AI not as a passive tool but as an “actant” within networks of researchers, institutions, journals, and knowledge practices. This approach reveals how AI becomes entangled with, and helps reshape, the norms, values, and power relations that govern academic knowledge production. Recent scholarship emphasizes that AI integration represents not merely a methodological shift but “a profound transformation in the epistemic structure of science” (Cruz-Aguilar, 2025).

Education research faces distinctive challenges here. Unlike fields focused primarily on technical precision, education research serves multiple audiences (teachers, students, policymakers, communities) and carries direct consequences for vulnerable populations. When AI becomes involved in producing educational knowledge, questions of accountability and representation take on heightened significance (Bearman et al., 2023). Moreover, education researchers have normative commitments to equity, justice, and democratic participation that shape how they understand their epistemic responsibilities.

The study will analyze how international bodies, publishers, and professional associations are currently constructing AI's role in scholarly work. What assumptions underpin their guidelines? Where do tensions or contradictions appear? Most importantly, how do these institutional responses either enable or constrain education researchers' ability to work thoughtfully and ethically with AI?

Three contributions are expected. First, the study will develop a typology of human-AI collaboration modes, distinguishing, for example, between using AI for language polishing versus conceptual development versus data analysis, and clarify their different implications for authorship and responsibility. Second, it will propose how originality might be reconceptualized: not as creation from nothing, but as distinctive synthesis, contextual judgment, and ethically grounded application. Third, it will identify emerging tensions in trust between transparency demands and practical limits, between individual responsibility and systemic challenges, that are particularly acute in education research, where trust relationships extend beyond academic peers to include practitioners and communities.

In alignment with the ECER 2026 theme, this study positions education researchers as responsible stewards of knowledge integrity in an evolving human-machine ecology, capable of leading thoughtful governance of emerging epistemic technologies.

This ongoing study critically analyzes the construction of digital sovereignty regarding the integration of Artificial Intelligence (AI) into the Portuguese education system. Drawing on an analysis of normative and regulatory documents across international, European, and national scales, the research challenges the technological determinism of inevitable modernization (Xiao & Bozkurt, 2025). Instead, the incorporation of AI is examined as a political and infrastructural problem that reconfigures power relations, redistributes public authority, and redefines modes of governance in the educational field (Gulson et al., 2022; Bratton, 2015). Consequently, digital sovereignty emerges not merely as a legal attribute, but as a simultaneously educational, political, and infrastructural issue co-produced across multiple scales of governance.

Despite its political centrality, digital sovereignty remains under-operationalized in education, constituting a missing link in the specialized literature (Parcerisa et al., 2024). This lacuna is critical as public schooling becomes a strategic site for datafication and economic value extraction, reconfiguring the material conditions of teaching through corporate infrastructures. The Portuguese case constitutes fertile analytical terrain: situated within a European framework marked by strong regulatory ambition yet persistent industrial fragilities, Portugal evidences the tensions between the discursive affirmation of strategic autonomy and the material dependency on exogenous technological ecosystems to integrate AI in education (Portugal, 2025a; Roberts et al., 2021).

Against this background, and considering the co-production of these dynamics across different scales of governance, the investigation is guided by the following research question: how do the normative and regulatory frameworks for Artificial Intelligence, articulated across international, European, and national scales, configure digital sovereignty within the Portuguese education system?

To address this question, the study adopts a sociotechnical and relational theoretical-conceptual perspective on digital sovereignty, understood here not merely as legal authority but as material and infrastructural capacity, produced through concrete practices, technical choices, and institutional arrangements (Belli, 2022; Musiani, 2022). Conceptually, the notion of ‘infrastructuring digital sovereignty’ is mobilized, enabling the observation of how ostensibly technical decisions (such as the adoption of public cloud services, global standards, or proprietary algorithmic models) perform specific forms of dependency and self-determination (Musiani, 2022). From this vantage point, sovereignty ceases to be conceived as a stable attribute of the State and becomes a contingent practice, continuously negotiated within asymmetrical relations with transnational private actors.

In articulation with this perspective, the theory of ‘organized hypocrisy’ illuminates the coexistence of normative discourses of sovereignty with administrative practices that naturalize technological dependency (Krasner, 1999; Santaniello, 2025). This lens exposes policies affirming strategic autonomy while delegating material control of educational infrastructures to global providers. At the confluence of political economy and governance, the study mobilizes ‘synthetic governance’ to examine the fusion of human decision-making and computational automation mediating pedagogical and administrative processes (Gulson et al., 2022). Consequently, AI integration is analyzed as a process of assetization, converting educational data into economic assets and placing the public mandate of schooling under tension (Komljenovic et al., 2025; Williamson et al., 2022).

Finally, the European and international relevance of the study lies in the articulated analysis of global (UNESCO), regional (European Union), and national documents, evidencing how transnational agendas of AI and digital sovereignty are translated, recontextualized, and materialized in the Portuguese case. In doing so, the study aims to contribute to the European debate by demonstrating that regulation, although central, proves insufficient to guarantee technological self-determination when dissociated from material control over educational infrastructures, thereby opening space for comparative dialogue with other European contexts marked by similar regulatory ambitions and convergent infrastructural fragilities.

The increased impact of neoliberalism in education has touched many aspects of the politics of education. Standardisation, quality, measurement, and effectiveness have supplanted the core expectations of education, such as social justice, the encouragement of self-actualisation, and the empowerment of individuals. While the need for data production in education contributes to monitoring progress and identifying children’s needs, it also reproduces power-related inequalities (Kelly, 2025; Bradbury and Robert-Holmes, 2017).

Early Childhood Education and Care (ECEC) has gained significant attention as a social investment model with diverse functions. These include preparing children for school, promoting their well-being, fostering social integration, enhancing employability, reducing social inequalities, and increasing women’s participation in the workforce (Alexiadou, Hjelmer, Laiho, and Pihlaja, 2024). In early years education, child-centred or holistic approaches have gradually supplanted data-oriented approaches. School readiness has become a focus of early childhood policy in England.

Ofsted (the Office for Standards in Education), alongside increased data collection, monitoring, and accountability within the EYFS, contributed to a performativity culture in the primary schools (Roberts-Holmes & Bradbury, 2017). Technology-supported, league-based platforms such as Decuypere, Grimaldi, and Landri have subsidised the marketisation of education and reduced the core of education to a consumer/customer-centred focus (Kelly, 2025). While data is positioned as a means of supporting children’s learning, research has raised concerns about its implications for professional autonomy, children’s rights, and everyday practice.

The whole neoliberalism-centred transformation of education brought long-term ramification of parent’s and children’s rights. Increased power-dominated datafication has left parents with no right to opt their children out of data collection. From a Foucauldian perspective, this school readiness concept and the datafication of children can be understood as forms of regulation of the future population and the deployment of biopolitical mechanisms to optimise human capital (Foucault, 1980, 2003).

This study aims to explore how data-driven policy expectations are interpreted by educators and families. This paper critically examines how school readiness is conceptualised and enacted within the English Early Years Foundation Stage (EYFS). It addresses how school readiness is constructed in EYFS policy, and how practitioners and parents interpret and respond to these constructions, and the impact of datafication and performativity culture on children’s rights.

Internationally, data and legislation have been given an increasing role in the governance of schools. However, little is known about how the increasing use of data and legislation affects roles, hierarchies, and educational practices in early years education (Bradbury & Roberts-Holmes, 2018; Rosén et al., 2021). In Norway, stateprovided mapping tests (implemented in 2004) and legislation on intensive training in early years education (implemented in 2018) was combined to ensure the priority of students in danger of falling behind in reading and numeracy in the system.

The schools and teachers have high autonomy when it comes to how data should be used and to how intensive training should be practiced, and we know little about what is actually done in this regard in the Norwegian municipalities and schools (Ministry of Education and Research, 2019–2020). In the recontextualization of a policy to becoming a pedagogical practice, ideology comes into play, and it is therefore difficult to foresee its realization(s) (Bernstein, 2000). What is important in relation to the recontextualizing of the legislation on intensive training is that both standardized data and legislation represent forms of governance that may furnish the state selected agents outside the school, as well as parents, with a stronger voice in the recontextualization process.

This paper examines how parents of students identified as falling behind have experienced the ways in which data derived from standardized testing, together with the statutory provisions on intensive instruction, have shaped power struggles over pedagogic priorities.

Theoretical framework

To analyze how power and control function in the recontextualizing of the legislation on intensive training into a pedagogical practice, the concepts of “recognition” and “realization rules” are employed. Recognition and realization rules are what establish the communicative context (Bernstein, 1990, pp. 34–35). Recognition rules refer to power relations and are regulated by the classificatory principle. Classification can refer to relations between contexts, agents, discourses, or practices in the intensive training and indicates how one context differs from another and “provides the key to the distinguishing feature of a context, and so orients the speaker to what is expected, what is legitimate to that” (Bernstein, 2000, p. 17). Framing refers to control over the communication, where an invisible pedagogy indicates that the acquirer has apparently much control over the communication (invisible pedagogy), and strong framing means that the transmitter has explicit control (visible pedagogy). The differences between the two forms of pedagogic orientations have been related to an ideological conflict between the new and old middle class over forms of control (Bernstein, 2000). Thus, the parents´ socioeconomic background could be relevant for how they experience power struggles and pedagogic priorities.