FAIR February 2026: Session “FAIR-Giving And FAIR-Taking”
All good things come in fours: like the four FAIR principles and the fourth Text+ FAIR February Workshop Session!
This year, the FAIR principles are celebrating their 10th anniversary. Since their publication by Mark D. Wilkinson and others in March 2016 as “The FAIR Guiding Principles for scientific data management and stewardship” in the journal Scientific Data, they have gained a fixed place within the requirements for digital research projects. But the principles focus on machine-readability of data, and they do not provide regulations for specific technical solutions, standards or specifications. Communities are called upon to find their own workflows and best practices.
This is where our FAIR February workshop series comes into play, aiming to discuss the implications and relevance of the FAIR principles for digital editions. Since the first workshop in 2023, the main topics have been: making editions and their research data findable; diverse aspects of interoperability; open licenses and conditions of use; and how to document one’s own work in a way which is helpful for data reuse. We also talked repeatedly about improving how accessible (i.e. barrierefrei) digital editions and their resources are, which is an important feature not included in the FAIR principles.
To mark the anniversary of the FAIR principles, this year’s program was organized in two sessions: “FAIR-nehmen und FAIR-geben” (“FAIR-Giving and FAIR-Taking”) and “FAIR-netzen” (“FAIR-Linking”). In the following, we will discuss the most important outcomes of the first session.
FAIR February 2026: Session One

For our session “FAIR-nehmen und FAIR-geben,” we organized a lightning panel discussion on the reuse of research data. We approached reuse as a process involving two complementary movements: preparing data so that it can be reused and accessing data that has already been made reusable. The discussion addressed both the challenges and potential solutions related to making research data reusable, and its subsequent reuse in research.
To discuss these questions, we invited four researchers whose work engages with different aspects of research data reuse and infrastructure. Carolin Odebrecht, a research assistant at the Department of German Studies and Linguistics at Humboldt University of Berlin, has worked on the European Literary Text Collection (ELTeC). Currently, she is involved in research data management at said department and at the Interdisciplinary Centre Digitality and Digital Methods. Elena Spadini, assistant professor at the University of Bern, works on research infrastructures, metadata, and the accessibility of research data. She recently co-organized the workshop “Re-Use of Editions and Text Collections: The Role of APIs” at the University of Zurich. Peer Trilcke, professor of 19th-century literature at the University of Potsdam and director of the Theodor Fontane Archive, has been closely involved in building digital humanities infrastructures, among them the Potsdam Network for Digital Humanities, the Journal of Computational Literary Studies, and the DraCor platform. Sven Jaros, a historian of Eastern Europe at Martin Luther University Halle-Wittenberg, recently coordinated the project “Modeling Pre-Modern Ambiguities (VAMOD)”, which explored best practices for making historical research data (such as charters, places, and personal names) available in the graph database FactGrid in accordance with the FAIR principles.
The lightning panel started with a short statement from each speaker on six different aspects of data reuse. The panelists had one to two minutes per question to share their perspective before the floor was opened for a ten-minute discussion with the audience. The fast-paced format kept the debate dynamic and engaging. What were the main takeaway points? What should we keep discussing in the future? You can find the answers to these questions in the following summary.
1. It’s working already! (?) – Which effects of data reuse in and from digital editions can already be observed?
In response to the first question, the panelists discussed current manifestations of data reuse in digital editions, asking whether such reuse already takes place and which practices enable—or hinder—it. Sven Jaros noted that digital editions are often developed as isolated projects, whose idiosyncratic structures can make reuse difficult.
Peer Trilcke approached the issue from the perspective of contemporary cultures of sampling and recycling. Referring to the Environmental Humanities Hackathon (Potsdam, 2023), where the digital poet JR Carpenter created a media assemblage combining texts by Carroll and Darwin with data from the International Comprehensive Ocean-Atmosphere Data Set, he raised the question of what such practices might mean for scholarly editions. His key point was that when data is available, it leads to practices of unforeseen reuse. This question is closely tied to the FAIR principles of visibility and findability.
Elena Spadini connected the discussion to the workshop “Re-Use of Editions and Text Collections: The Role of APIs,” which had taken place the week before. She highlighted ongoing work in digital editing and publishing and pointed to data aggregators as practical examples of infrastructures that already facilitate reuse, mentioning Historical Dictionaries and, most notably, CorrespSearch.
Carolin Odebrecht emphasized the importance of interdisciplinarity and argued for a clearer separation between project data and interface data. A case in point is the Text+ Cooperation Project “Pessoa digital project”, which modernized and restructured its technical infrastructure by separating the corpus data of the edition from the platform’s interface data and publishing them in two distinct GitHub repositories.
To sum it up, we can say that ‘it’s working’—although maybe in unexpected ways. Data aggregators play a crucial role in making edition data findable. However, there is still work to be done to foster open science practices, open up projects to other users (and uses), and to encourage greater interdisciplinary.
2. Who – how – what? – Where exactly does reuse take place in digital editions? Who is using data from digital editions? For which purposes?
When discussing examples of data reuse in digital editions, Sven Jaros mentioned the annotation of named entities, such as persons and place names, as a case in point. Most digital editions include such annotations, ideally shared in standardized formats that allow reuse across projects. One such exchange format is BEACON. Named entities are also often linked to authority files such as the GND or Wikidata, enabling connections between datasets and research infrastructures. Questions of data linking were also central to the second session of FAIR February 2026, “FAIR-netzen,” where BEACON and authority files played a major role. At the same time, Jaros noted that linking data to authority files raises questions of data quality, particularly when it comes to using community-maintained repositories like Wikidata.
Peer Trilcke reminded the audience that the most intensive users of data from digital editions may currently be bots harvesting content for the training of large language models. He referred to the considerable load such automated requests place on the digital services of the Theodor Fontane Archive, describing these actors as agents of “digital colonialism.” Such systems may eventually reproduce content derived from digital editions in their outputs, and this form of reuse raises serious political and environmental concerns, Trilcke argued.
Elena Spadini noted that digital editions are often primarily used by editors themselves who continuously work with the material. Editions also play an important role in teaching. At the same time, existing corpora are increasingly explored through computational approaches such as topic modelling or the study of intertextual relationships. As an example, Spadini referred to the Old Bailey Proceedings Online, which features a list of digital projects using Old Bailey online data; this is an initiative that demonstrates how showcasing specific cases of reuse can help communicate the possibilities of openly available data.
This point resonates with Carolin Odebrecht’s call for making data visible. She emphasized that making both software and data visible is a fundamental prerequisite for reuse. While this reflects a core FAIR principle, her remark also highlighted a persistent challenge: in practice, the findability of research data remains limited. Odebrecht therefore stressed the importance of addressing visibility and reuse already in the early phases of a research project, which ensures that the data produced later will meet the requirements for meaningful reuse.
We can safely say that reuse occurs across a wide range of areas in research and teaching. That being said, we already touched upon the challenges brought about by AI, whose training methods heavily rely on open access data. A core prerequisite for reuse is the findability of data, something we have to consider already in the planning stages of a project.
3. I’ve prepared something for that. – What does it take to ensure good reusability? What prevents it?
It was no surprise that findability remained a central aspect of the discussion. Elena Spadini argued that a crucial requirement for good reusability is not only that data can be found, but that it can also be accessed easily. There are many ways to achieve that, prominently placing a download button is one example. Offering an API that provides standardized data and enables flexible, machine-readable access is another method that significantly enhances reuse. This argument was backed by Peer Trilcke, who challenged the idea of “owning” data and instead advocated for a more open approach: “to let it go; spread it.” Trilcke argued that the quantity of user interaction is of greater value for data reuse than high-quality data that remains unused.
This view was challenged by one of the participants, Christian Thomas (Berlin-Brandenburg Academy of Sciences and Humanities, BBAW), who pointed out, that, in practice, beyond findability, data quality is indeed a decisive factor for reuse. Perhaps this is not an either/or question, but rather a case of “it depends,” namely on the intended purpose of use. Sven Jaros pointed out that although Wikidata and FactGrid have a somewhat chaotic, “grassroots” character—that is, they are shaped by decentralized, community-driven contributions—they can nevertheless foster exchanges with other users and projects that can be very inspiring for one’s own research.
Beyond the discussion around findability and data quality, the speakers emphasized other important aspects. Carolin Odebrecht, for example, underlined the importance of documentation and communication for a clear understanding of the material and possible (re)uses. In one sentence: “Document what you know—and what you don’t know.”
To achieve good reusability, this aspect has to take center stage: Provide easily findable and manageable access to data, release data as early as possible and talk about what you are doing with your data.
4. Are we there yet? – Where/When does reuse begin? At what stage of a project should the question of data reuse be addressed?
Carolin Odebrecht stressed that reuse begins at the very start of the research process. This is the moment when we have to reflect on what we want to learn from our data and, ultimately, what others might want to learn from it as well. Elena Spadini concurred by noting that it is necessary to plan ahead. She suggested starting with potential data users: who they might be, what they might want to do with the data, and what they would need in order to do so.
Peer Trilcke argued that planning should begin as early as possible, ideally already at the stage of project conception, echoing Carolin Odebrecht’s point. He then adopted a different approach to the issue by questioning the distinction between data use and reuse, suggesting that this distinction should be minimized. Project planning often prioritizes use and thus control over data and what ultimately happens to it. Reuse is often overlooked in the beginning, but because it extends far beyond the lifespan of a project (at least ideally) it should be a central concern from the outset. Thus, he asked: at what stage of a project are we prepared to relinquish control? He strongly advocated for “letting go” and for placing reuse at the core of research practices.
Sven Jaros pointed out that we still need to reflect on the question “what is data?” or “what is research data in the humanities”, since many resources and materials are not yet recognized as ‘data’. Carolin Odebrecht had raised this aspect earlier, arguing that there is a need for more theoretical work on qualitative metadata, datasets, and reuse processes that are inevitably shaped by human subjectivity. She underlined the importance of teaching students how to engage with these issues and to incorporate such reflections into their everyday research practice.
The takeaway message here was quite clear: reuse should not be treated as an ‘afterthought.’ Critically thinking about and planning reuse should accompany your project from the get-go. Good practice in this realm still lacks a robust body of theoretical work on what constitutes data and on different aspects of data quality. Producing this knowledge as a community and consistently integrating it early on in curricula is the key to achieving widespread good practice.
5. A FAIR question! – For which challenges of digital editions do the FAIR principles (not) offer solutions?
Carolin Odebrecht highlighted that the FAIR principles primarily focus on data, while questions concerning web applications remain insufficiently addressed. In the context of digital editions, however, web interfaces play a central role as the presentation layer between the data, the digital methods employed, and the users. Odebrecht argued that researchers should be seen not only as creators of research and providers of data, but also as service providers. As such, they should be concerned with the usability of their products.
Another aspect that is not addressed by the FAIR principles is data ethics, whose importance was highlighted by Elena Spadini. This is a perspective, however, which is articulated in the CARE Principles.
Peer Trilcke pointed out that copyright issues are also not addressed by the FAIR principles. He acknowledged that this was not their original purpose, but questions of data reuse involve broader sociopolitical dimensions that need to be considered. Within the framework of “open science,” data is often understood as a common good; however, this openness has also been exploited for economic gain within profit-driven contexts. Participant Barbara Fischer (German National Library, DNB) raised the question of how the obligation to share publicly funded research data relates to issues of economic fairness, asking whether those who profit from such data should also be held accountable, for example through taxation. Trilcke finally raised the question of whether the FAIR principles can be reclaimed and disentangled from their appropriation by policy agendas. Perhaps we should turn back to the very beginning and ask a more fundamental question: what were the original goals behind these principles?
Sven Jaros argued that the FAIR principles have not yet led to significant changes in research culture. At the same time, he sees great potential in interlinking data from different projects that would previously not have been considered related. As an example, he mentioned the use of place names collected from the database of the Arolsen Archive in a project concerned with medieval ruling practices in Eastern Europe. This example illustrates how standardized, linkable data can move beyond the boundaries of individual projects and even historical periods, thus creating unexpected connections.
The FAIR principles have improved the interoperability of data and the exchange within the research community, and digital editions do profit from this. However, this seems not to be enough to be truly FAIR: there are issues of data ethics, copyright, research culture, and good practices for web services that challenge digital edition workflows, research, planning, and reuse. We need to address these challenges as well.
6. AI does that already. – In what ways can AI support the generation of reusable edition data? How does AI impact the criteria for ensuring data reusability? And are new forms of responsibility emerging in response to AI-induced change?
Ultimately, the discussion around this question can be summed up in one phrase: Responsibility is key.
The first arguments were in favour of AI. Sven Jaros pointed out that AI can be helpful regarding often time-consuming processes such as data mapping and enrichment. Elena Spadini also pointed to the potential of using AI for specific tasks. She noted that AI can be particularly useful for documentation, which is often needed but frequently neglected. In addition, AI can support interaction with machine-readable datasets (for example, with RDF data through RAG-based applications) by enabling access via natural language. In this sense, AI can serve as a means of exposing and presenting data to users.
On the other hand, Spadini emphasized the need for caution, raising the question of how to conceptualize data literacy in the context of AI. She addressed the downsides of AI, pointing to a growing discrepancy between the “easy-to-use” experience of AI chatbots and the complex power structures behind them. She warned that open digital methods may be at risk in a landscape increasingly shaped by AI monopolies and emphasized that these challenges cannot be addressed by researchers alone but require political action. However, she also stressed the responsibility within academia to strengthen digital literacy.
Peer Trilcke concurred with this cautious perspective, remarking pointedly that “these wonderful AI tools are evil,” and noted the ecological and environmental impact of AI infrastructures. Just like Spadini, he called for a responsible approach to AI in research, and advocated for smaller, locally hosted models, public infrastructure, and, more broadly, a resource-efficient, open, and diversity-conscious approach to the use of AI in editing and data curation. While he acknowledged that it remains uncertain whether the scholarly community can meet these challenges, he emphasized that practices of digital editing are likely to change radically in the emerging AI paradigm.
Carolin Odebrecht also emphasized the importance of critically and carefully threading the needle. She highlighted the central role of data and digital literacy. “We need to know what we are doing,” she noted, referring to her own institution, Humboldt University in Berlin, which provides a locally hosted AI model for safe use in teaching and research.
Odebrecht summarized three key points for engaging with AI: first, greater responsibility, both in how we use AI and in providing good examples of its use; second, the need for literacies, teaching why and how to use these tools; and third, the need for more theory, particularly with regard to developing a stronger theoretical foundation for AI methods.
AI can be a very useful tool for research, analysis, and documentation. There is certainly potential that AI will radically change digital editing. It presents new possibilities as well as new challenges, but it may be threatening the open science culture as we know it. So, yes, new responsibilities are emerging. In order to counter the growing exploitation by AI monopolists the scientific community needs to develop responsible ways and methods for handling and using AI and, most importantly, teaching students these aspects as part of digital literacy.
There is a lot we can do as a community, but AI has become a societal issue that needs to be discussed by a larger audience. It is important that the developments in AI are understood as a political issue that affects research, funding, education and society as a whole.
Conclusion
One participant stated in the closing discussion: “It’s all about communication!” From what you have just read, it should be clear: There is still a lot to talk about! The following key points came up several times throughout the discussion: we need to make resources findable, plan thoroughly and start early with the documentation, and, of course, communicate what we are doing.
Nur der Text ist unter der Lizenz Creative Commons Namensnennung 4.0 International nutzbar. Alle anderen Elemente (Abbildungen, importierte Anhänge) sind „Alle Rechte vorbehalten“, sofern nicht anders angegeben.
OpenEdition schlägt Ihnen vor, diesen Beitrag wie folgt zu zitieren:
Sandra König, Fernanda Wolff (21. Mai 2026). FAIR February 2026: Session “FAIR-Giving And FAIR-Taking”. Text+ Blog. Abgerufen am 10. Juni 2026 von https://doi.org/10.58079/1694p

