Who is actually the author when artificial intelligence is involved in the creation of Open Educational Resources (OER)? This question is rapidly gaining importance. With the proliferation of generative AI tools, the boundaries between human creativity, automated text and image production, and collaborative work processes are blurring more than ever. OER traditionally stand for transparency, sharing, and collaborative creation. But what happens when part of that creation comes from a machine?
Written by Shahla Rasulzade
When Creation Becomes a Shared Process
A teacher sits down to adapt an open textbook. Instead of starting from scratch, she turns to a generative AI tool. Within seconds, it produces explanations, examples, even exercises. She reviews, edits, restructures – and publishes the final version as an Open Educational Resource (OER). At first glance, the process seems efficient, even empowering. Yet a fundamental question quietly emerges: Who is the author of this resource?
This question is no longer hypothetical. As generative AI becomes embedded in educational practices, it challenges one of the most deeply rooted assumptions of OER: that knowledge is created and owned by humans. OER has long relied on a clear and principled understanding of authorship. Creators, adapters, and remixers are recognized through open licenses that emphasize attribution, transparency, and reuse (UNESCO, 2012; David Wiley, 2021).
. This framework reflects a human-centered model of knowledge production – one in which contributions are visible, traceable, and accountable. However, this clarity begins to dissolve when AI enters the creative process.
Generative AI and the Rise of Hybrid Creation
Generative AI systems are no longer passive tools. They actively participate in shaping content by generating initial drafts, restructuring arguments, translating materials, adapting content for different audiences. In many OER workflows, creation is no longer a linear human activity but a collaborative interaction between human intention and machine generation. This gives rise to a new form of production: Hybrid authorship – a space where content is neither fully human nor fully machine-generated, but co-produced. At the center of this transformation lies a simple question: What counts as authorship when AI is involved?
Legal Perspectives: A Fragmented Landscape
In practice, contributions exist along ranging from minimal human intervention to deeply iterative collaboration. However, current systems tend to treat authorship as binary: either human or not. Across jurisdictions, copyright law continues to privilege human authorship (World Intellectual Property Organization, 2019). For example, in the United States, works lacking substantial human input are typically ineligible for authorship (Thaler v. Perlmutter, 2023). In the European Union, authorship is closely tied to human intellectual and moral rights (P. Bernt Hugenholtz & João Quintais, 2021; EU Copyright Directive, 2019). China adopts a pragmatic regulatory approach in which AI-generated content can receive protection when supported by clear human input, while state oversight ensures lawful data use, algorithm registration, and alignment with broader industrial and governance objectives (Tencent v. Yingxun, 2019). In the United Kingdom, authorship may be attributed to those who arranged the creation process. Despite these variations, one principle remains consistent: AI is not recognized as an author. For globally shared OER, this legal fragmentation introduces uncertainty that directly affects attribution, licensing, and reuse.
Invisible Contributions and Ethical Blind Spots
Beyond users and developers, generative AI introduces another layer of complexity: the invisible contributors. AI systems are trained on vast collections of human-created content – books, articles, images, and educational materials. These sources shape AI outputs, yet their creators are rarely acknowledged (World Intellectual Property Organization, 2019). This raises important questions: Are AI-assisted OER indirectly built on uncredited human labor? Should attribution extend beyond visible contributors? In this sense, AI does not only blur authorship – it challenges the very boundaries of whose contributions are recognized. Since OER is not merely about access, but also about equity, transparency, and trust, this issue challenges its principles too. If AI contributions remain opaque, the integrity of educational resources may be questioned. Moreover, without mechanisms to track contributions, authorship disputes become more likely and licensing decisions become increasingly complex. The integration of AI risks undermining the very principles that define OER – unless new approaches are developed.
Towards Transparent Hybrid Authorship
Rather than forcing AI into existing categories, a more productive approach is to rethink authorship as a multi-layered, hybrid process, which acknowledges both human and machine contributions and distinguishes between different types of input (generation, editing, structuring). It may make the creation process more visible and transparent. AI itself may play a role in enabling this transparency by analyzing differences across OER versions and supporting semantic versioning systems. Such approaches could transform OER from static resources into traceable knowledge ecosystems, where contributions are not only recognized but also understood.
Rethinking Authorship: From Ownership to Contribution
The emergence of generative AI invites a conceptual shift from “Who owns this?” to “Who contributed, and how?”. This shift becomes especially relevant in AI-enhanced OER, where human and machine contributions are increasingly intertwined.
In this context, distinguishing between different types of contributions becomes essential. Kullmann and Rasulzade (2025) propose a useful lens by differentiating between content-neutral and content-based changes. While content-neutral modifications – such as formatting, grammar correction, or minor edits, often supported by AI – typically do not constitute authorship, content-based contributions – such as adding new material, restructuring explanations, or adapting content for learning-reflect deeper intellectual input. Applied to AI-supported creation, this distinction helps clarify a critical point: not all AI-assisted work carries the same authorship value. While AI may efficiently support surface-level improvements, meaningful authorship remains tied to human contributions that shape the substance, structure, and pedagogical intent of the resource. In this sense, humans remain central – not as sole creators, but as those who guide, interpret, and take responsibility for the knowledge being produced.
Conclusion: Designing Fair and Transparent OER Futures
Generative AI is not just a technological development – it is a conceptual turning point. It challenges long-standing assumptions about creativity, authorship, and ownership. For OER, this challenge is also an opportunity. By rethinking authorship through the lens of transparency and fairness, OER communities can lead the way in shaping responsible AI integration. The question is no longer whether AI should be part of open education. The question is how we ensure that its role remains visible, accountable, and aligned with the values of openness.
References
Kullmann, S., & Rasulzade, S. (2025). What Is a Recognisable Contribution? On the Characteristics of OER Authorship.
EU AI Act (2024). Artificial Intelligence Act.
Thaler v. Perlmutter (2023). United States District Court for the District of Columbia. Retrieved from WIPO Lex.
Tencent v. Yingxun (2019). Technology Co., Ltd. Beijing Internet Court. Retrieved from WIPO Lex.
UNESCO (2012). Paris OER Declaration.
Wiley, D. (2021). The 5Rs of Open Educational Resources.
World Intellectual Property Organization (2019). WIPO Technology Trends: Artificial Intelligence.
Hugenholtz, P. B., & Quintais, J. P. (2021). Copyright and Artificial Creation: Does EU Copyright Law Protect AI-Assisted Output?
Dieser Text steht unter der CC BY 4.0-Lizenz. Der Name des Urhebers soll bei einer Weiterverwendung wie folgt genannt werden: Shahla Rasulzade für OERinfo – Die Informationsstelle OER
