Conference Agenda

As part of the international open science movement, the openness of research data has emerged as a goal both to enable transparency and reproducibility of research and to support the reuse of datasets (Pampel & Dallmeier‑Tiessen, 2014). Research organizations, funders, and publishers promote the opening of research data in various ways, and a wide range of guidelines and services are available for this purpose. Despite requirements and guidelines for openness, research data are still opened only to a limited extent, and several challenges remain (Chauvette et al., 2019; Mozersky et al., 2022; Shero et al., 2025). Open research data are also reused relatively little in research (Gend & Zuiderwijk, 2022), leaving the potential of open data underexploited. There are differences between disciplines and data types in this respect. While in some fields the practices for opening and reusing data are already established, certain types of data are considered more challenging, and no established models or expertise for opening them have developed. The research culture in education science is often filled with concerns regarding the sharing of data, such as protecting your research work and participants, even though there are solutions for these concerns, at least for quantitative data (Logan et al. 2021). For example, fear of incorrectly deidentifying data or accidentally including private information is one identified barrier to data sharing in the fields of education and psychological research (Shero et al., 2025). In qualitative research, various epistemological, methodological, legal, and ethical aspects pose challenges for opening research data (Chauvette et al., 2019). Qualitative datasets are therefore seldom shared openly, but the pressure to open them has nevertheless increased (Mozersky et al., 2022). Opening data collected from human participants involves ethical challenges. Both participants and researchers may consider the data confidential for various reasons. Researchers are often left to deal with the ethical challenges of openness on their own. Especially in these changing times of poly-crisis, it is important that research data can be effectively reused while at the same time ensuring the safety of researchers and research participants.

This study explores how artificial intelligence (AI) can support open research data practices in school leadership studies, with particular attention to ethical, participatory, and governance-related challenges. While Open Science promotes transparency, accessibility, and reusability of research data, educational leadership research often relies on highly contextual, sensitive, and relational data involving school principals, teachers, and institutions. The integration of AI into data preparation and reuse processes introduces new opportunities, but also raises concerns regarding data sovereignty, consent, power relations, and trust.

The central research question guiding this study is:

How can AI-supported data preparation, anonymisation, documentation, and reuse practices enhance Open Research Data in school leadership studies while maintaining ethical integrity and participatory governance?

By addressing this question, the study aims to move beyond technical discussions of AI and instead focus on everyday research practices, researcher responsibilities, and governance arrangements in educational leadership research.

Theoretical Framework

The theoretical framework of this study is grounded in an interdisciplinary combination of Open Science, participatory research, and democratic educational governance, with AI positioned as a mediating socio-technical actor rather than a neutral tool.

First, Open Science provides the overarching normative framework. Open Research Data is understood not only as a technical requirement for data sharing, but as a practice aimed at increasing transparency, traceability, and cumulative knowledge production in education. From this perspective, data quality, documentation, and reusability are closely linked to accountability and public trust in educational research—issues that are particularly salient in comparative and cross-national European research contexts.

Second, the study draws on participatory research and shared governance approaches in educational administration. These perspectives emphasise that research participants—such as school leaders and teachers—are not merely data providers but stakeholders with legitimate interests in how data are interpreted, shared, and reused. This lens is crucial when considering open data practices, as openness can unintentionally reproduce power asymmetries if participants are excluded from decisions about data governance. Integrating participatory principles into AI-supported data practices allows the study to examine how consent, agency, and co-ownership of data can be operationalised in Open Science.

Third, the framework incorporates democratic educational governance, focusing on how open and reusable leadership data can support evidence-informed decision-making, institutional learning, and policy development. In a European and international context, comparable and reusable data on school leadership practices can contribute to cross-country learning, policy dialogue, and capacity building. At the same time, governance-oriented perspectives highlight the risks of decontextualisation and misuse of leadership data when detached from local educational realities.

Within this framework, AI is conceptualised as a boundary object that connects researchers, infrastructures, and data users across national and institutional contexts. AI-supported processes such as anonymisation, metadata generation, and data curation are examined critically in terms of transparency, accountability, and inclusiveness. Rather than positioning AI as a solution in itself, the study treats it as a catalyst that amplifies existing ethical and governance questions in Open Research Data practices.

By combining these perspectives, the study contributes to international debates on how Open Science can be implemented in educational leadership research in ways that are ethically robust, participatory, and aligned with democratic values across European and global research communities.

This presentation will focus on the problematization of semantic data structuring through the lens of situated curating in two digital platforms, ParVO and ecHoimages, both of which focus on research on arts education. Arts education is not merely an application field for these platforms, but an epistimological zone that spans, for instance, cultural mediation, teaching, design, and artistic practices. So the platforms are built, problematized, and inscribed in this framework. Both platforms are Wikibased (https://parvo.wikibase.cloud/ and https://echoimages.labs.wikimedia.pt/) and are being tested with students in the arts field. From this experience, we developed the concept of decyclopedization [1] and explored situated knowledge [2] as an archival possibility [3], [4], [5], [6]. Right now, we mobilise the concept of decyclopeding as a way of (un)learning [4], [7], using subversive strategies and tactics within the Wikimedia ecosystem to achieve alternative and marginal knowledge. We also strive to promote educational exercises and challenges that incorporate error, playfulness, and fiction into the dynamics of research and pedagogy with these tools. These dynamics allowed us to question assumptions of neutrality and universality embedded in dominant data models and encyclopedic platforms, and to reflect on how relational, partial, and context-dependent forms of knowing can inform the conception of data models. It also highlights how digital systems classify, relate, and legitimize knowledge. Rather than sharing results, we expose the relational logics through which data are organized, allowing others to reconfigure the ways a problem is investigated.

More recently, we have been developing the idea of situated curating [8], [9], as a way of resignifying heritage from a fixed object of preservation to a living, relational, and contested process, grounded in the investigation of ruins as spaces for artistic practices and curatorial actions. In this sense, these spaces are seen not only as architectural remains, but also through the entanglements between surrounding nature, human communities, and more-than-human inhabitants. Situated curating challenges the way we think of arts education and the respective platforms for teaching and research. In this sense, what kinds of platforms and data models should be developed for arts education in the context of situated curating?

We mobilize situated curating here as a reference for understanding how positionality and political concerns inform data practices. In both cases, knowledge emerges from following relations rather than from fitting phenomena into predefined categories. So it becomes concrete in everyday decisions about how data are entered, described, and connected in the platforms. While data is usually employed as a technical task, our project approaches it as a curatorial process in which decisions about entities, properties, and relations shape what becomes visible within a cataloguing system.

This approach requires new perspectives and languages that question traditional methods of archiving as well as broader regimes of knowledge organization and production. Regarding data collection and cataloging, this poses an enormous challenge, as situated and context-dependent forms of knowledge cannot be fully accommodated by fixed, universal categories, requiring a restructuring of semantic data through new vocabularies, models, and ontologies. It also becomes important when the data are intended to be shared and reused by others, a central concern in open research and educational data practices. In this regard, we seek the FAIR (Findable, Accessible, Interoperable, Reusable) [10] and CARE (Collective Benefit, Authority to Control, Responsibility, Ethics) [11] frameworks.

Complex transdisciplinarity and epistemological intersections require new knowledge graphs (KG) assisted interfaces that enable research across heterogeneous systems and evolving research questions, treating data as capta [12] by supporting narrative, relational, and exploratory modes of inquiry. In this sense, KG supports exploration across relations, not just record storage. Our platforms promote visual querying by graphs and spacetime views.

Research data are increasingly positioned as central to transparency, cumulative knowledge production, and innovation in educational research, yet the everyday practices through which data are prepared, shared, and reused remain uneven and contested (Raffaghelli, Battistig, et al., 2025; van der Zee & Reich, 2018). This tension is particularly evident in research contexts where data are relational, sensitive, and ethically charged, such as educational research involving learners, institutions, and vulnerable populations (Hernández-Leo et al., 2023). While Responsible Research and Innovation, embedding Open Science (Owen et al., 2012), provides a powerful normative framework promoting openness, transparency, and reuse, less is known about how these principles are interpreted and enacted in practice by researchers, especially at early career stages (Dai et al., 2018; De-Filippo et al., 2023; Raffaghelli & Manca, 2022). Addressing this gap, this study examines Open Research Data practices through the lens of doctoral researchers, combining a research-based synthesis of prior studies and project evidence with an empirical investigation based on a questionnaire administered within doctoral training workshops on Open Data practices between 2023–2024 and 2025. The questionnaire was iteratively developed and validated through a two-round Delphi study involving international experts in educational research, digital scholarship, and data literacy, ensuring conceptual coherence, construct validity, and practical feasibility (Raffaghelli, 2019).

Furthermore, the study is grounded in a theoretical framework that conceptualises research as a situated, socio-technical activity embedded in contested and problematic digital environments (Jandrić et al., 2024) under transformation, from being open, digital and networked (Veletsianos & Kimmons, 2012) to becoming datafied and platformised (Raffaghelli, Ferrarelli, et al., 2025). Particularly, constitutive elements of scholarly work require dynamic practices of data seeking, generation, processing, sharing, evaluation, and reuse as the base to produce or validate knowledge (Dutton, 2011). This perspective is enriched by scholarship on digital and networked academic practices, which foreground openness, participation, and collaboration, while also acknowledging the infrastructural and political conditions that shape contemporary research within datafied and platformised digital instruments and environments. Within this framework, data practices are thus treated not as neutral technical procedures, but as ethically and institutionally mediated actions that contribute to researchers’ professional identities.

The findings reveal that doctoral researchers largely endorse the values of openness, transparency, and public accountability associated with Open Science, and they report frequent engagement in generating and managing their own research data, often using established tools and producing traditional forms of data representation. At the same time, practices related to data sharing, reuse, and publication in open repositories remain sporadic and marginal. Notably, ethical reflection emerges as a salient and comparatively robust dimension of research activity: respondents report frequent consideration of ethical issues related to data extraction, treatment, and presentation, often exceeding their engagement with formalised frameworks for data quality assessment or open data standards. Importantly, these patterns point to a form of situated data literacy shaped by disciplinary norms, professional socialisation, and asymmetric institutional support structures. Data preparation and sharing emerge as negotiated practices involving judgment, care, and risk assessment, particularly in relation to qualitative data, research involving minors or vulnerable groups, and the secondary use of data by third parties or AI systems whose training data and analytical processes often remain opaque. Learning about data practices is largely sustained through informal and self-directed learning ecologies, relying on peer relationships and self-selected resources rather than formal institutional provision.

Overall, this contribution speaks directly to ongoing debates within Network 12 on the practical challenges of data sharing and reuse in educational research.