What Do Knowledge Graphs Bring to OA Publishing – A Reflection
Authors: Ursula Rabar and Julien Homo
In recent years, knowledge graphs have become a frequent topic of conversation, particularly for their role in making research more open, more navigable, and more coherent. Within the GRAPHIA project, we are working to build the first comprehensive Knowledge Graph for the Social Sciences and Humanities (SSH), enriched with Artificial Intelligence and Large Language Model applications. With this infrastructure, GRAPHIA aims to enhance data visualisation and analysis, help researchers uncover patterns in unstructured data, and illuminate social and cultural phenomena with an unprecedented level of clarity.
But how does this emerging capability intersect with the publishing ecosystem in all its formats and outputs?
Because the data we draw on is publicly available, integrating published content (such as open access journal articles, books, chapters, and related outputs) into a knowledge graph brings clear benefits. These advantages span the entire publishing supply chain, from authors and editors to librarians, publishers, and platforms. While each group may engage differently, many benefits overlap: improved discovery, richer connections between entities, easier identification of relevant research, authors, topics, funding, publishers, or datasets, and new opportunities for reuse.
Looking at just two points (authors and publishers) in the OA books supply chain already reveals how impactful this can be.
For Authors
A knowledge graph offers authors a powerful single entry point into the research landscape. Rather than searching across multiple platforms, they can quickly discover relevant literature, identify potential collaborators, and locate associated datasets or related research outputs. Their own work also becomes more visible and more accurately indexed, thanks to the graph’s structured connections between topics, institutions, publications, and contributors. Using a Knowledge Graph helps reduce the noise in exploration and research, while at the same time it expands an author’s ability to explore, position, and disseminate their research.
For Publishers
Publishers experience many of the same advantages. A connected view of the scholarly ecosystem allows them to identify emerging authors and topics, understand where new research areas are growing, and see how their own publications sit within wider disciplinary networks. This helps inform commissioning strategies, editorial decisions, and community engagement. Knowledge graphs also enhance metadata workflows by linking to authoritative sources and exposing relationships that support reviewer selection, marketing, and catalogue development. Ultimately, they offer publishers a richer, more dynamic understanding of their content and its place in the broader research environment.
The Artificial Intelligence Aspect
Knowledge Graphs also help address a growing concern in scholarly publishing: the impact of AI and large language models. Generic LLMs often hallucinate, invent references or misrepresent authors and works because they lack a stable, authoritative semantic structure. A Knowledge Graph provides exactly this foundation, a curated network of authors, publications, topics, institutions and their verified relationships. When AI tools are grounded in such a graph, they no longer guess: they retrieve facts from authoritative linked data, which reduces hallucinations, improves citation integrity, and ensures that identifiers and relationships remain consistent.
It also enables more trustworthy generative services, such as summarisation, topic extraction or metadata enhancement, because every assertion can be traced back to a known source with documented provenance. In this sense, integrating a Knowledge Graph into the publishing workflow is not only a way to improve discovery and metadata quality, it is also a way to make AI safer, more explainable and better aligned with the values of open scholarship, particularly within infrastructures like GRAPHIA and GoTriple.
How Are We Moving Forward?
Of course, adopting knowledge graphs requires some learning: becoming familiar with their functionalities, understanding how they can be queried, and integrating them into existing workflows. It may also feel like “one more platform” at first. Yet having a single, connected access point to a rich body of structured information more than justifies the effort. The benefits to discoverability, efficiency, metadata quality, and strategic insight are substantial, and they grow over time as more entities and relationships are added.
As knowledge graphs become more embedded in research infrastructures like GRAPHIA, their value to authors, publishers, and the broader scholarly communications ecosystem will only increase. Within our work, we have been, and still are, looking at use cases and user needs that can help bring more clarity to the value of the GRAPHIA Knowledge Graph but also how it should be developed to make sure it fits the requirements of its users. Further reading can be done here. As the project progresses, we will be sharing more information and results to make the value proposition for the OA publishing supply chain clearer.
If you wish to be involved or have any questions, you can contact us at contact@graphia-ssh.eu as well as subscribe to our newsletter.