AI learning landscapes as metaphors: Expanding horizons of academic reflection through collaborative creative inquiry linked to the UNESCO draft Dubai Declaration on OER

Chrissi Nerantzi¹, Margaret D. Korosec¹, Yasmin Kader¹ and Sandra Abegglen²

¹ University of Leeds, Leeds, UK
² University of Calgary, Calgary, Canada

Abstract

This paper uses UNESCO’s draft Dubai Declaration on OER (Open Educational Resources) as a focal point for critical and creative reflection. The authors examine the text to consider what stands out, the ideas and dilemmas it provokes, and how these can be expressed through a combination of written and visual language, using metaphors as a form of creative inquiry. This approach enables richer and more diverse interpretations, revealing insights that might otherwise remain hidden.

A reflective narrative in the form of dialogue – interwoven with visual elements – creates space for a plurality of authentic voices among the co-authors and co-researchers. The study identifies three key themes – Transparency, Translation, and Teamwork – offering a deeper understanding of OERs, Artificial Intelligence (AI), and their implications for education. By integrating visuals and metaphors throughout, the paper bridges abstract concepts with tangible interpretations, fostering a richer and more inclusive exploration of the subject matter.

Keywords

metaphors, GenAI, OER, Dubai Declaration, creative inquiry, photographs, dialogue

Background

This paper was written at the time citizens of the world were invited to comment on UNESCO’s (2024) draft Dubai Declaration on OER. We saw this as an opportunity to come together, three scholars from the University of Leeds and members of the Knowledge Equity Network, and a scholar from the University of Calgary. We advocate for open and inclusive education to discuss and articulate our individual and collective responses linked to specific areas more deeply. UNESCO defines open education through Open Educational Resources as learning, teaching, and research materials in any format and medium that are in the public domain or released under an open license. Although the final Dubai Declaration on OER has been released with some changes, the following discussion regarding openness, technology, and education remains pertinent. This dialogue continues to support the practical implementation of the final document, particularly in relation to leveraging digital public goods and emerging technologies to enable equitable and inclusive access to knowledge. We are sharing these hoping to engage others in a wider conversation to broaden our collective understanding and propose our ideas to be considered to further strengthen the work on AI (Artificial Intelligence), OER and education where this may be appropriate towards the democratisation of and equitable access to knowledge.

Artificial Intelligence (AI) has been an explosive cocktail, a real dynamite for academia dominating higher education conversations, conferences, and publications while practices largely remain the same. There are of course the experimenters, the rebels, those who are driven by their curiosity to break free from conventions and make new discoveries to stimulate learning and change for the better. They have not stayed away from AI. On the contrary, they have actively experimented from the outset. Many of them are open scholars and have shared their experiments more widely (such examples include Abegglen et al., 2024; Nerantzi et al., 2023).

Yet, in the corridors of universities and other higher education institutions, discussions around AI usually centre around assessment and academic integrity – alarm bells are ringing. Students are accused of cheating, lecturers are scrambling to adapt their teaching practices, and institutions are grappling with how to uphold academic standards in an era where AI challenges traditional notions of originality and authorship. True? While open scholars raise concerns about the origins of data, lack of transparency, sustainability, and the profiteering of a few from the knowledge of many, they also recognise and embrace the opportunities that these changes bring. Are we clinging to notions of ownership – often over ideas that stem from collective contributions – or are we creatively and openly seizing the new possibilities before us?

This paper approaches the discussion on AI and education creatively, using visual metaphorical thinking, based on an already existing, collectively developed and discussed, declaration on OERs.

Methods and methodologies

Collaborative autoethnography in dialogic form using reflection was used by the co-authors to inquire together into their individual and collective understanding and connection to the UNESCO’s (2024) draft Dubai Declaration on OER in relation to open education and AI. Collaborative autoethnography provides the vehicle to explore authentically diverse voices and gain deeper insights into these as expressed by the individuals themselves (Blalock & Akehi, 2017; Ellis et al., 2011; Lapadat, 2017). The individuals who came together for this inquiry are colleagues across different higher education institutions, open scholars, with some of them members of the Knowledge Equity Network with an interest in the intersection of open education and AI.

The co-authors as collaborative, situated researchers, fully aware of their positionality as co-authors and co-researchers, respond in this study to the following research question:

What thoughts were triggered by engaging collectively with the draft UNESCO (2024) Dubai Declaration on OER and what are the implications for open educational practices?

Data was collected via Google docs using a dialogic reflective approach (Bohm, 2004) using written and visual language combined with metaphors to achieve greater richness and diversity, and to make sense of our own emerging and more established thoughts. Botha (2009) states that metaphors – beyond being a useful linguistic device in literature and commonly used there – also play an important role in education as they enable diverse and richer expression of thoughts, emotions, ideas, and reflections as well as interpretations of experiences. We use metaphors to make sense of new ideas and experiences by referring to the new ideas and experiences using the known and familiar. Geary (2011, p. 13) states characteristically “A metaphor is both detour and a destination, a digression that gets to the point”. Each individual approaches this uniquely, using various metaphors to describe the same phenomenon, thought, or idea, with none being inherently better or worse, right or wrong (Botha, 2009). In our case, photographs were employed to illustrate personal emotional connections and to visualise the metaphors used during the conversation, helping to articulate and further elaborate on specific positions or ideas (Hartel, 2016). Therefore, the analysis and interpretation of the photographs and metaphors are recognised as personal – and individual (Botha, 2009; Richardson, 2000); this allows for deep collaborative autoethnographic insights.

Co-authors and co-researchers are referred to in the reflective conversation as Author 1 (CN), Author 2 (MDK), Author 3 (YK), and Author 4 (SA). The conversation is presented as it originally unfolded to maintain individual voices and authenticity.

Reflective conversation

From our collective reading of the draft UNESCO (2024) draft Dubai Declaration we identified three areas to explore through an authentic reflective dialogue weaving in personal visual metaphors:

1. Transparency

2. Translation

3. Teamwork

These themes emerged through studying the draft Declaration and synthesising areas explored within this. We used the themes as conceptual lenses to conduct the conversation which follows. This will help us make sense of our own positionality, thoughts, and ideas on OERs and AI, and also gain insights into the authentic perspectives of the other co-authors and co-researchers in this study:

1. Transparency

Context: The draft Declaration emphasises active transparency throughout all phases of AI development, including training data, to ensure trustworthiness and accountability. It calls for open, replicable AI systems based on open-source software and openly licensed content, alongside measures to protect privacy and prevent monetisation of educational data. Transparency is linked to provenance and attribution, with recommendations for cryptographic signing, metadata embedding, and regular audits of the OER ecosystem to build trust among stakeholders.

Author 1: It was really good to see in the draft declaration transparency as fundamental to AI in relation to open education. There does not seem to be much transparency at the moment in relation to the data that is used by Large Language Models (LLMs) and Generative Artificial Intelligence (GenAI) applications while there is wider call for it (Urmeneta & Romero, 2024).

When I thought about transparency and what this means to me in relation to AI, I remembered an experience and a picture I took a few years ago when I was on Corfu at the beach (see Appendix under 1.1). It was early morning, nice and warm, not too hot. No noise, just a few other people and I could hear the tiny waves splashing on the sand. After making myself comfortable, I decided to go for a swim. I stood there with my toes in the water and looked around me. I could see my toes, the sand, and tiny fish swimming around. I loved that. I looked around, over the horizon. The endless blue. If the water was clear here, just imagine how it was over there? I continued walking, after I left my phone under my umbrella on the beach. I started swimming. I moved my arms and legs and looked around. I could still see what was in the sea. I felt so calm. It was really soothing. This is what transparency means to me. How about everybody else?

Author 3: I love this delightful example of your experience at Corfu beach! It was so vivid that I could imagine myself being there too – and indeed wanted to be there with you!

For me, an experience that comes to mind when I reflect on transparency in the GenAI context relates to my very first visit to a stunning, new university in Doha, Qatar. It was a warm, sunny day in Doha with a calm, blue sky and clear, bright sunlight cheering us. As my taxi drove away from busy downtown Doha and out towards the expansive, beautiful, new Education City region, my driver chatted away volubly. After some time, he suddenly pointed out two magnificent ‘spires’ of the university Minaretein building towering on the horizon. This first glimpse took my breath away. The gleaming white minars looked like elegant beacons – a stunning visual landmark (see Appendix, under 1.3).

It was clear that the university’s distinctive design drew both on its culture and natural environment whilst being modernistic and futuristic. The sense of openness and cultural connection was represented not only in its exposed physical structure but also in its explicit decorative style; for example, exquisite calligraphy – advocating the relationship between knowledge and light – was integrated into the very fabric of the huge surrounding walls of the central courtyard.

And there was a sense of light upon light everywhere. That is to say, shafts of natural daylight pierced through multiple, pretty openings in the ceilings, walls, corridors and doors, penetrating beyond physical layers to offer further visibility and insight. Indeed, the bright sunlight seemed to reach the spaces above, below, across and beyond the building itself, connecting nature with the learners, academics, and community. Creating oneness.

The university term had not yet commenced (and the workmen were still pottering away with the final touches) but nevertheless its openness could be seen in the varied diversity of the people already using the building. Members of the local community were welcomed and encouraged to use its learning areas, libraries, cafes, exhibitions, and gardens. Families and children were there, enjoying themselves.

But for me, the sense of openness was most discernible at the university’s spectacular prayer hall where this serene, expansive space enabled people to individually reflect, meditate, think critically, recite, and engage in dialogue together. The sense of community, equity and peace was profound as eminent academics and business leaders humbly stood shoulder to shoulder alongside diverse students, staff, cleaners, construction workers, families, and children.

So, my experience of this building and of what it offered as a learning space for all people is an example of what transparency and openness means for me.

Author 1: This is indeed fascinating. For me it was the natural surroundings and for you a man-made structure in the landscape. I hope we will visit that beach together one day. Spaces seem to evoke emotions and help us reflect and make sense of our thoughts. Do they also provide valuable framing and help us to notice insignificances that are actually significant? Just wondering.

Author 4: I am struck by how the notion of transparency resonates in both of your accounts, Author 1 and Author 3, particularly in how it intertwines with nature and the built environment – the surroundings and contexts that shape our lives. There’s definitely something there that helps us notice insignificances that are actually significant when speaking about openness and AI.

Transparency evokes, for me, the image of glassware: clear and see-through, yet undeniably present and tangible (see Appendix under 1.4). Glass serves as a powerful metaphor for something simultaneously strong and fragile, resilient yet delicate.

Reflecting on the UNESCO’s (2024) draft Dubai Declaration on OER, I am drawn to its emphasis on “active transparency” across all phases and aspects of AI development. This phrasing suggests more than mere openness or visibility; it implies intentionality, a dynamic and ongoing commitment to clarity and accountability. Active transparency could also be interpreted as a deliberate effort to not only reveal but also engage with the processes, decisions, and structures underlying GenAI systems.

Returning to the metaphor of glass, active transparency might mean acknowledging both its beauty and its fragility. It reminds us that while transparency can illuminate and connect, it also requires care and attentiveness to preserve its integrity. Just as glass can distort or shatter under pressure, transparency without thoughtfulness can mislead or fracture trust. It is this balance – of strength, delicacy, and purpose that we must strive for in AI and beyond.

Author 2: These are wonderful descriptions of transparency. I love both the similarities and differences in the natural and man-made spaces.

For me, transparency means being able to see what is intended to be seen. In human-designed spaces, including learning resources, learning environment, and experiences, it refers to encountering something in the way its designer hoped it would be perceived. Yet this ideal is always informed by learners’ diverse backgrounds, prior experiences, and perspectives. No matter how carefully we design, we cannot fully orchestrate how an experience will be received. Perhaps this is, in fact, the essence of openness: not being constrained by any single, intended experience. It is an interesting paradox – we design learning opportunities, interactions, activities, and assessments with particular outcomes in mind, yet we release control over how they are interpreted and enacted.

I appreciate how Author 4 uses glass as a metaphor for transparency. It invites us to ask whether colours or textures of glass represent forms of active transparency. I can clean my windows to maintain clarity – that is within my control. But what about glass that is opaque, textured, stained, or dependent on the presence of light? A window is designed to allow us to see outward, to view the landscape beyond. What if the glass itself is a mosaic of colours and patterns (see Appendix under 1.2)? The expectation of seeing through becomes complicated by the nuances of what we are asked to see through. Active transparency, then, introduces a dynamic and contingent framing, whereby transparency is not simply a state, but a set of practices shaped by context, material conditions, and intentional actions.

I am particularly struck by the relationship between intentionality and agency, as outlined by Author 4. Malle et al. (2001, p. 3) describe intentionality as a “quality of actions” in contrast to intention as an individual’s “mental state that represents such actions.” This distinction is provocative when considering the use of AI. Could AI be regarded as an intentional partner in learning? And if so, does it have any form of agency that would allow it to act as such? These boundaries force us to rethink what participation means in digitally enhanced learning.

UNESCO’s (2024) draft Dubai Declaration on OER positions the need for active transparency across the entire lifecycle of AI development, alongside commitments to protect data from monetisation. This is the intention of the declaration. Yet the intentionality – the actual quality of actions – undertaken by those who create, share, and reuse OER, as well as those who develop and deploy AI, is still outside of any single actor’s control. The declaration recognises this tension but cannot fully resolve it; instead, it challenges us to navigate transparency, agency, and openness with care.

Author 1: Those in power and access to the data sell it often off for profit without consent, without agreement to feed the hungry data monsters. How can this be right?

Author 2: Exactly – we feed the data monsters through our participation in open systems. We tick “I agree” with hope, trust, and belief that our information will be protected, yet the power imbalance between individual users and large corporate actors makes this trust challenging.

As open practitioners and scholars, we are driven by values of generosity, reciprocity, and knowledge sharing. We want to give freely and allow others to reuse, remix, and repurpose our work. This ethos directly confronts the business models that depend on capturing, analysing, and monetising data – well beyond what might be considered reasonable consent.

I am not confident we will ever reach agreement with large corporations to use exclusively openly licensed data in training their systems. Such a shift would demand a fundamental reimagining of copyright itself. It would require a seismic transformation in how we conceptualise ownership, authorship, and the public good in digital spaces.

Author 4: The data monster… insatiable, ever hungry, and endlessly greedy. It makes me question whether changes to copyright and policy are truly enough to tame its appetite, particularly when we consider the concept of transparency. What kind of transparent practices could effectively keep the data monster at bay – if any? Or does the solution lie elsewhere, perhaps in the realm of ethical AI practices? Would we need a binding Code of Ethics, one that enforces accountability and fairness, to truly address the challenges posed by its relentless consumption? Transparency alone might not suffice; it may require a deeper, systemic shift toward a more ethical and principled framework for AI development and data governance – a framework like the one begun by the Dubai Declaration on OER?

Summary: This conversation indicates that we see transparency as fundamental to AI in open education, yet we acknowledge that current practices lack clarity around data use and system design. Our conversation highlights ethical concerns about ownership and monetisation, and we use metaphors like clear water, glass, and architectural openness to express transparency as both strong and fragile. For us, “active transparency” means more than visibility. It is an intentional, ongoing effort to ensure fairness and trust. We also question whether transparency alone can address systemic issues without deeper ethical frameworks.

2. Translation

Context: Translation is framed as a key enabler of inclusivity and equity in the draft Declaration. It advocates for AI-based translation services to make OER accessible in multiple languages, while stressing the importance of contextualisation and collaboration with local communities to maintain cultural relevance. It warns against pure linguistic conversion and promotes quality assurance in AI-generated translations, ensuring outputs respect diversity and avoid reinforcing systemic imbalances.

Author 1: The Declaration suggests that text-to-speech or AI-based translation services can support activities to make OER more accessible. This made me wonder if it is enough to be able to translate quickly using GenAI based on what data sets exactly? At the moment we are in the dark! We know that most of the data sets will be from Western countries and in English (Bozkurt et al., 2024). The same applies to OER (Karakaya & Karakaya, 2020; Lambert, 2018).

The origin of the data sets that are used by large corporations are unknown to the end-user, also to the originators, the creators (Bozkurt et al., 2024).

Author 2: When I think about translation, I initially think about making information available to someone in their own language when they are unable to understand the original. It can also refer to converting complex concepts into simpler forms for easier consumption. A visual metaphor related to navigation compares a globe with a map. Although a globe offers a view of the earth, local maps are necessary for detailed guidance (see Appendix under 2.2). Local maps offer specific context, directions, and options for selecting the best route. Sometimes, established routes do not require additional assistance for navigation. I am fascinated by the historical Silk Road, which crossed continents, climates, and cultures. It served as an important channel for economic development and linguistic diversity. I wonder if everyone travelled with a translator or if the common language was one of exchange and barter. What would OER look like if translation were unnecessary and access was universally clear?

Author 1: It also matters where the outputs generated will be used and how. Is just translating data sets and GenAI outputs in different languages enough or even appropriate? Regarding OER this has been recognised as insufficient as this means the local context and culture are ignored (Karakaya & Karakaya, 2020; Lambert, 2018).

When I started thinking about translation and providing ‘fruits’ to others, the photo I took of a pot on a wall high up in a greenhouse in the Botanical Gardens in Cambridge came to mind with a hanging plant (see Appendix under 2.1). I can’t remember its name. The plant reminded me of Rapunzel’s hair, something like a ladder but also the saying ‘low bearing fruits’. But is this enough? Does this create dependency? In the context of OER and AI, we create, we translate for you – and you use? On one side the creators and givers? And on the other side the users? Does this sound right?

Author 3: I agree. Important as they undoubtedly are, it is not enough to only be able to translate quickly from English, or to largely draw on data sets from the West, especially given their opaque origins or indeed, for largely OER ‘fruits’ from the West to be given to others, even when efforts are made to localise them. Such practices are valuable, but they may still be perceived as patronising and perpetuating a dependency culture. For me, this situation raises important and contentious questions of ethics, power, and equity. Fiendishly difficult as these issues are, they need to be acknowledged, and their implications explored and addressed, particularly when we know that open education is quintessentially about promoting social justice (UNESCO, 2019); and that Sustainable Development Goal 4 (United Nations, 2015) calls on the international community to ensure inclusive and equitable quality education and to promote lifelong learning opportunities for all.

Another reflection on the Declaration’s ambitions for Translation is how we ‘translate’ these important intentions into reality on the ground and specifically regarding the role governments should play in creating enabling and supportive environments for institutions and others to help deliver this agenda.

Of course, a plethora of (open) education activities are already taking place, but all too often the issues relating to systemic, structural power, and resource allocation imbalances in education are ignored or remain unaddressed. Leadership by our governments for this OER agenda is hugely significant. As a former policy adviser, I appreciate how elusive engaging governmental decision makers can be, but if we are to create structural, sustainable, transformative change within our institutions, then we need the support of our governments. More could be achieved, and more quickly, if the declaration’s objectives were part of our collective national vision and priorities for education.

Hence, regarding translation, I suggest that part of our efforts could usefully include exploring ways to connect with governmental decision-makers to encourage their active ownership of the educational social justice agenda and thus help translate these ambitions into meaningful actions. That would help us progress initiatives within our own institutions to help rebalance our teaching and learning contexts and collaborate meaningfully with relevant stakeholders, ensuring we include, co-create, and empower marginalised others.

Author 2: I agree that empowering others is essential for a sustainable, high-quality, and accessible OER ecosystem. I value the Declaration’s emphasis on collaboration with communities to ensure relevant translations.

Author 1: How can we overcome the limitations of translation? What localisation? As a translator for over 10 years, I understand the need for localisation so that the output produced in the source language is also relevant in the target language. Without this filter, this adaptation, I am not sure that there is a lot of value in just generating for example resources using GenAI for a completely different culture and/or context. Ok, so we say translation is not enough as data sets used in generating AI outputs will not be relevant to in the target culture or context. How can the GenAI resources then be localised? And what about the need for creating OER in local languages (Lambert, 2018)? In local languages? How can this be enabled through partnership working for example? And how if we know that OER update of materials created in English still have a low update?

Author 3: I think those of us in more wealthier parts of the world might usefully explore efforts to overcome the digital divide through meaningful partnership with marginalised, under-represented and underserved groups. For example, by reaching out, including and actively listening to their voices, we can discover how together we might explore more creative, innovative ways to support them. We could perhaps help build their capacity and skills and help provide access to resources, tools and systems so that they themselves can better create content, curate resources, update existing materials, co-create and facilitate collaborative teaching, learning and research on their own terms, in their own languages and relevant to their own priorities, needs and contexts. Perhaps such approaches might help address some of the challenges of achieving effective, faithful interpretation, translation, and ‘localisation’.

Author 2: These last comments deeply resonate for me. I agree that engaging communities for cultural context in translation is important, but supporting local language OER contributions may be even more effective for global visibility. However, this still risks echoing a dynamic where Western cultures intervene to assist marginalised voices.

Author 1: We talk a lot about decolonisation, Western culture and the English language dominating the discourse. English dominates the OER landscape. While using English as a lingua franca has been recognised as an important connector, the higher-level English often used in OER creation is not understood more widely or at all by non-English speakers. This creates marginalisation and exclusion and results in widening educational inequalities (Karakaya & Karakaya, 2020; Lambert, 2018).

What role is GenAI now playing in all this? Is it potentially leading to more rapid domination of western ideas that are shared in English at scale? Urmeneta and Romero (2024, p. 30): “We must […] avoid the reductionist trap of designing didactic tools detached from the context in which they are used”. What does this really mean in the context of generating resources for learning?

Author 3: Yes, these are very real challenges. Certainly, given the global dominance of English as the lingua franca for storing and sharing human knowledge across disciplines and boundaries and for business, science, technology, and culture, it's unlikely that things will change any time soon. Therefore, diverse peoples across the world wanting to benefit from AI are still required to secure an adequate level of English. And we know too that AI powered translation tools are becoming increasingly sophisticated and better at addressing cultural and linguistic barriers. Nevertheless, AI systems still lack the nuanced understanding and intentionality that reflect human language use. So, though language models are picking up patterns based on the data that they’re trained on, they’re still not able to learn meaning in the rich refined manner humans do. It reminds me a little of the story of The Sorcerer's Apprentice, where the young student sneaks a peek at his master’s powerful book of magic and tries to use a spell to avoid him having to mop the floor. The results are catastrophic. He subsequently learns not only that there are no shortcuts to hard work but that powerful tools require great care and skill to utilise effectively. And I think this is the point about GenAI: it's a valuable, powerful tool that needs to be handled with caution and criticality. Hence, we need a wider range of people to for example, feed into the data sets and to be providing the necessary ‘prompts’, namely the nuanced inputs and questions to help create relevance, to contextualise and to reduce bias.

Author 1: When we start exploring GenAI tools we instantly see freemium and premium versions. A highway and a scenic path perhaps? How can this be good for equity? How is this making knowledge available for all? In the Declaration there seems to be a lot about ‘content' but we have acknowledged for some time now that education is not about content. Why do we keep referring to learning and teaching materials as ‘content’? Isn’t ‘context’ far more relevant and what will make the real difference?

Author 4: I agree that context is crucial, particularly when discussing translation. There’s more to translation than simply converting content from one language to another, no matter how accurate the translation may be. Context provides the nuance and depth necessary to truly understand and convey meaning. When we relate this to AI, it becomes especially clear that we must move beyond the simplistic global versus local dichotomy – the translation from one to the other (see Appendix under 2.4). We exist simultaneously as citizens of the world, members of our local communities, and participants in online spaces. Within this complex interplay, I resonate with Author 3’s observation that “all too often, issues relating to systemic, structural power imbalances are ignored or remain unchanged”. This highlights that translation alone is insufficient; it must be accompanied by a deeper awareness of context and a commitment to addressing these power dynamics.

Author 3: Yes, I agree. Ultimately, greater equity will be secured when the diverse peoples of the world have equity of access, the necessary skills, and can themselves produce/adapt contextualised learning resources in relevant languages. Of course, achieving such a vision will take time, but determined small, steps can help advance this journey. I think greater investment in such systemic equity and inclusion efforts by the more privileged stakeholders and co-created with under-served others, could help in progressing improvements.

Summary: We understand translation as more than converting words; it is about cultural and contextual adaptation to avoid reinforcing inequities. We question whether rapid AI-driven translation, often based on Western and English-centric datasets, is enough. For us, localisation and co-creation with communities are essential to ensure relevance and empowerment, rather than creating dependency. Through metaphors like maps, globes, and fruits, we explore the complexity of navigating global and local needs. We connect translation to systemic power imbalances and advocate for strategies that enable marginalised groups to create and adapt resources in their own languages.

3. Teamwork/Collaboration

Context: Collaboration is presented as essential for implementing the UNESCO (2019) OER Recommendation and addressing emerging challenges. The Declaration calls for multi-stakeholder engagement, including educators, learners, policymakers, technologists, and legal experts, at local, regional, and international levels. It promotes participatory governance, co-creation of OER using AI platforms, and the development of collaborative frameworks for ethical AI use, provenance tracking, and capacity building. This collective approach aims to foster equity, sustainability, and innovation in the global OER ecosystem.

Author 1: Yes, to collaboration beyond an institution, geographical areas, and cultures. Also collaborating with others outside higher education. The benefits are multiple as they diversify learning and teaching, bring in novel perspectives and stretch and challenge our thinking and practice; they also help us come closer together as human beings (Nerantzi, 2018; Nerantzi, 2019).

But what also about collaboration within institutions? Open scholars often feel lonely, unheard, and disconnected from their own institutions (Nerantzi, et al., 2025). A fish out of water comes to mind on a sandy beach (see Appendix under 3.1). Lying there on its own, the little fish, unable to move, unable to breathe. As open educators we very much value collaboration. GenAI from a user perspective, seems to be a human to machine collaboration. Can I say collaboration? I think it is that talking to each other, working together, learning together that we value as humans. Now we can learn with the machine? What does this mean for human-to-human collaboration?

Author 4: Definitely yes to human collaboration and co-creation! Human collaboration is fundamental (Abegglen et al., 2021; Abegglen et al., 2023). However, collaboration with machines – specifically the software driving AI – is something we are still trying to navigate. Like Author 1, I question whether we can truly talk about human-AI collaboration, or can we? Some students might argue that AI serves as a ‘teacher’ or ‘study buddy’ (Abegglen et al., 2024), but is AI genuinely a partner? Or is it more akin to a mirror – reflecting back what we project onto it, shaped by its programming and our interactions, but ultimately devoid of genuine agency or reciprocity? (see Appendix under 3.4).

In this rapidly evolving landscape, how do we uphold the value of human collaboration? How can we continue to work together, professionally, and individually, without losing ourselves in the process or feeling like a fish out of water, particularly when striving to remain open to new possibilities, with or without AI?

Author 3: I absolutely agree with the need for collaboration within and beyond our institutions. It's important to try and find like-minded individuals and to gradually build up a supportive, dynamic coalitions for equity and social justice via open education. Every individual and their contribution - or simply their ambitions – matter and can be built upon.

Author 2: Collaboration is an essential part of learning and interacting within our society. The individuals and institutions we interact with impact the opportunities that are possible. I really value how Stodd et.al. (2023) framed the human capacity for learning as a process of formal and informal experiences which leads to sense making activity, including how we interact with artificial intelligence. As we engage in sense making activity with technology, we also need to look at the opportunities for engaging more with the future of work and society.

I value Author 4’s mirror comparison and the idea of reflection. I remember a day out in London some years ago. While walking, I observed the London Eye reflected in a puddle. The water was transparent and still, allowing for a distinct reflection of the iconic structure (see Appendix under 3.2). If a pebble is thrown into the puddle or if it begins to rain, the image becomes distorted or unclear. Visibility depends on environmental conditions, similar to how teaching and learning are influenced by human actions and decisions (Ferreira et al., 2023). Reflections can shift quickly. Therefore, verifying information, consulting sources, and reviewing assumptions are important processes.

Reflection highlights the need to critically evaluate what we see, how we learn, and how we engage and collaborate – with humans or AI. Perhaps the issue extends beyond collaboration between humans and AI, encompassing our metacognitive understanding of our interactions with AI, strategies for community development, and efforts to foster a collaborative network of colleagues. This consideration also prompts further examination of whether AI can serve as an effective partner, working collaboratively to support our learning. Dede (2024, p. 1) considered an extension of the role of GenAI as a “knowledgeable partner in teaching, learning, and assessment” framing IA – intelligence augmentation – as a “complementary partnership”. Can our teamwork and way of collaborating with others allow us to see through and look behind as we collectively tackle the wicked problems we are facing?

Author 1: What policies, practices, and cultures do we need to create equitable places and spaces to nurture and recognise their contributions fully and respectfully?

Author 4: And, to add to that question, how do we ensure these policies, practices, and cultures are not just performative but deeply embedded and sustainable? How do we move beyond surface-level inclusion to create environments where contributions are truly valued? In other words, what mechanisms can we put in place to actively dismantle systemic barriers and power imbalances while fostering a sense of belonging and mutual respect – for all contributors?

Author 2: Institutional culture plays a role in nourishing a thriving community of open education practitioners and scholars. Open education may remain peripheral to core institutional activities unless it is integrated into professional development frameworks and recognised as a legitimate path for career advancement. We also need to check our assumptions related to “degrees of openness” as not all institutions have clear policies or guidance on open practice and scholarship (Patel et al., 2023, p. 293). Active open education practitioners and scholars are driven by the desire to advance knowledge and act as agents of change by purposefully contributing to openly accessible resources and research. However, when an institution’s strategic objectives do not align with the principles of openness, strategic dissonance may arise, potentially undermining the efforts of individual change facilitators (Korosec, 2017).

Author 1: There is also a clash between the commercial or for-profit business models of large AI corporations and what open education stands for, which is social innovation and social enterprise to change the world.

Author 2: I agree. The hope of open is to make a difference in the world and to spark curiosity and elicit interest in others. The essence is co-creation and collaboration. This can be done through creative and innovative ways of engaging as a dispersed community. Urmeneta and Romero (2024, p. 10) note that “Education is not merely about imparting knowledge; it is about nurturing creative thinking, fostering critical skills, and empowering individuals to become lifelong learners”.

Learning design continues to be vital in promoting creativity and innovation within open education. While this remains a key focus, there is a growing need to intentionally establish a solid foundation for creativity and thoughtfully integrate the role of Generative AI in the open education environment. Leveraging GenAI to support creative approaches will help guide progress and shape the future of open education. Creativity is fundamentally associated with generative AI, and it remains an area for ongoing exploration and study (Atenas et al., 2024).

Industry and start-up companies ought to be included in the creative process. This is potentially mutually beneficial for educational institutions and industry. Fundamental in this approach is the inclusion of students. The student perspective is so important! I’ve seen the impact active collaboration has on students. Their confidence soars as they learn they have something valuable to contribute. This also extends the reach of learning about AI as a solo endeavour to one of collaboration and community.

With the speed of change and ambitions in developing and deploying digital public goods, as outlined in the Declaration, there is also an urgency to develop digital literacy, and specifically AI literacy. Approaching this through practice-based, experiential, and collaborative ways helps build connection and community. We know the power of grass roots efforts on the ground with change facilitators and committed individuals building momentum. To ensure professional development at scale, digital and AI literacy must be supported strategically at an institutional level.

Author 4: I completely agree that digital literacy, including AI literacy, should be rooted in practice-based and collaborative approaches. Emphasising co-creation and collaboration aligns perfectly with the principles of openness and innovation.

The challenge lies in transforming the current education system and AI technologies into a more inclusive and dynamic learning ecosystem – one that empowers individuals and groups to collaborate, innovate, and create meaningfully together. How can we ensure that this transformation is guided by human ethics, respecting values such as equity, transparency, and social responsibility?

Summary: The conversation reveals that we value teamwork as vital for equity and innovation, extending beyond institutional and cultural boundaries to include diverse stakeholders such as industry and students. We reflect on the loneliness of open scholars within institutions and question whether AI can truly be a collaborative partner, often seeing it as a mirror that reflects rather than reciprocates. Human-to-human collaboration remains central for us, but we recognise the need for metacognitive awareness in how we engage with AI. Using metaphors like fish out of water and reflections in puddles, we highlight vulnerability and the contingent nature of clarity in collaboration. We call for systemic changes in institutional culture, policies, and practices to embed openness and nurture inclusive partnerships.

Concluding remarks

In conclusion, this paper leverages the UNESCO (2024) draft Dubai Declaration on OER as a lens for critical and creative reflection, emphasising the importance of engaging with concepts in ways that allow for nuanced, diverse, and often hidden truths to surface. By combining reflective narrative with visuals and metaphors, we foster a dialogue that is not only intellectually engaging but also creatively resonant, allowing for a deeper exploration of the themes and tensions inherent in the notions of openness and AI. The study identified and explored three key themes – Transparency, Translation, and Teamwork – offering a deeper understanding of OERs, AI, and their implications in education using a collaborative autoethnographic approach. This approach not only meshes the diverse voices of the authors but also demonstrates the potential of creative methodologies to enhance critical discourse.

The paper uses the declaration as a lens for creative and critical reflection on AI and openness in education, revealing insights into our three interrelated themes. Transparency emerges as a call for active, intentional clarity in AI development and data governance, challenging the opacity of current practices and highlighting ethical concerns around ownership and monetisation. Translation is framed not merely as linguistic conversion but as a culturally situated process requiring localisation and empowerment to avoid reinforcing systemic inequities and Western dominance. Teamwork underscores the enduring value of human collaboration while questioning whether AI can be a genuine partner, advocating for co-creation across institutional, cultural, and disciplinary boundaries to foster equity, creativity, and AI literacy, recognising the important that governments have in creating enabling environments?

Through metaphors and visuals, this collective autoethnographic study illustrates how creative inquiry surfaces nuanced tensions, between openness and control, global and local, human and machine, while proposing that inclusive, context-sensitive practices and ethical frameworks are essential for shaping equitable futures in education. This suggests that for OER to fully realise this vision in an AI-mediated world, future practice must be guided by creative, context-sensitive, and ethically attuned frameworks that sustain both human and technological forms of knowing.

Declarations

Data availability

The data used in this paper is made available within the paper and the Appendix due to the methods used. The photographs are all the authors own.

Competing interests

The authors have no competing interests to declare.

Authors' contributions

Chrissi Nerantzi made substantial contributions to the conception or design of the work; made substantial contribution to the acquisition, analysis, and interpretation of data for the work; drafting the work, revising the draft critically for important intellectual content

Margaret Korosec made substantial contribution to the acquisition, analysis, and interpretation of data for the work; drafting the work, revising the draft critically for important intellectual content

Yasmin Kader made substantial contribution to the acquisition, analysis, and interpretation of data for the work; drafting the work, revising the draft critically for important intellectual content.

Sandra Abegglen made substantial contribution to the acquisition, analysis, and interpretation of data for the work; drafting the work, revising the draft critically for important intellectual content

References

Abegglen, S., Burns, T., & Sinfield, S. (Eds.) (2023). Higher education collaboration: A new ecology of practice. Bloomsbury.

Abegglen, S., Burns, T., & Sinfield, S. (Eds.) (2021). Collaboration in higher education: Partnering with students, colleagues and external stakeholders. Journal of University Teaching and Learning Practice, Special Issue, 18(7). https://open-publishing.org/journals/index.php/jutlp/issue/view/61

Abegglen, S., Nerantzi, C., Martínez-Arboleda, Karatsiori, M., Atenas, J., & Rowell, C. (Eds.) (2024). Towards AI literacy: 101+ Creative and critical practices, perspectives and purposes. #creativeHE. https://doi.org/10.5281/zenodo.11613520

Atenas, J., Nerantzi, C., & Havemann, L. (2024). Critical and creative pedagogies for AI and data literacy: An epistemic data justice approach for academic practice. Journal Research in Learning Technology, 32.

Blalock, A. E., & Akehi, M. (2017). Collaborative autoethnography as a pathway for transformative learning. Journal of Transformative Education, 16(2), 89-107. https://doi.org/10.1177/1541344617715711

Bohm, D. (2004). On dialogue (2nd ed.). Routledge.

Botha, E. (2009). Why metaphor matters in education. South African Journal of Education, 29(4), 431-444. https://doi.org/10.15700/saje.v29n4a287

Bozkurt, A. Xiao, J., Farrow, R., Bai, J. Y. H., Nerantzi, C., Moore, S., Dron, J., Stracke, C. M., Singh, L., Crompton, H., Koutropoulos, A., Terentev, E., Pazurek, A., Nichols, M., Sidorkin, A. M., Costello, E., Watson, S., Mulligan, S., Honeychurch, S., Hodges, C. B., Sharples, M., Swindell, A., Frumin, I., Tlili, A., Slagter van Tryon, P. J., Bond, M., Bali, M., Leng, J., Zhang, K., Cukurova, M., Chiu, T. K. F., Lee, K., Hrastinski, S., Garcia, M. B., Sharma, R. C., Alexander, B., Zawacki-Richter, O., Huijser, H., Jandrić, P., Zheng, C., Shea, P., Duart, J. M., Themeli, C., Vorochkov, A., Sani-Bozkurt, S., Moore, R., & Asino, T. L. (2024). The manifesto for teaching and learning in a time of generative AI: A critical collective stance to better navigate the future. Open Praxis, 16(4), 487–513. https://doi.org/10.55982/openpraxis.16.4.777

Dede, C. (2024). Guest editorial Intelligence augmentation: The owl of Athena. IEEE Transactions on Learning Technologies, 17, 2100-2101. https://doi.org/10.1109/TLT.2024.3456072

Ellis, C., Adams, T. E., & Bochner, A. P. (2011). Autoethnography: An overview. Historical Social Research/Historische Sozialforschung, 36(4 (138)), 273–290. http://www.jstor.org/stable/23032294

Ferreira, G., Lemgruber, M., & Cabrera, T. (2023). From didachography to AI: Metaphors teaching is automated by. Journal of Interactive Media in Education, 2023(1), 1-13. https://doi.org/10.5334/jime.798

Geary, J. (2011). I is an other. The secret life of metaphors and how it shapes the way we see the world. Harper Prennial.

Hartel, J. (2016). Pictorial metaphors for information. Journal of Documentation, 72(5), 794-812. http://www.jennahartel.info/uploads/8/3/3/8/8338986/pictorial_metaphors_for_information.pdf

Karakaya, K., & Karakaya, O. (2020). Framing the role of English in OER from a social justice perspective: A critical lens on the (dis)empowerment of non-English speaking communities. Asian Journal of Distance Education, 15(2), 175-190. http://www.asianjde.com/ojs/index.php/AsianJDE/article/view/508

Korosec, M. D. (2017). A case study of change facilitators for technology-enhanced learning in relation to one university’s strategic intent [Thesis]. University of Hull. https://hull-repository.worktribe.com/output/4255572

Lambert, S. R. (2018). Changing our (dis)course: A distinctive social justice aligned definition of open education. Journal of Learning for Development, 5(3). https://doi.org/10.56059/jl4d.v5i3.290

Lapadat, J. C. (2017). Ethics in autoethnography and collaborative autoethnography. Qualitative Inquiry, 23(8), 589-603. https://doi.org/10.1177/1077800417704462

Malle, B. F., Moses, L. J., & Baldwin, D. A. (2001). Intentions and intentionality: Foundations of social cognition. Massachusetts Institute of Technology.

Nerantzi, C. (2019). The role of crossing boundaries in collaborative open learning in cross-institutional academic development. Research in Learning Technology, 27. https://doi.org/10.25304/rlt.v27.2111

Nerantzi, C. (2018). The design of an empirical cross-boundary collaborative open learning framework for cross-institutional academic development. Open Praxis, 10(4), 325-342. http://dx.doi.org/10.5944/openpraxis.10.4.907

Nerantzi, C., Abegglen, S., Karatsiori, M., & Martinez-Arboleda, A. (Eds.) (2023). 101 creative ideas to use AI in Education. #creativeHE. https://doi.org/10.5281/zenodo.8072949

Nerantzi, C., Kader, Y., & Korosec, M. (2025). Open scholars: Putting their heads above the parapet or transitioning from Grimm’s Rapunzel to Disney’s Tangled. Open Scholarship of Teaching and Learning Journal, 4(1).

Patel, D., Parsley, S., Cannell, P., and Havemann, L. (2023). Practicing open education. In L. Amrane-Cooper, D. Baume, S. Brown, S. Hatzipanagos, P. Powell, S. Sherman, & A. Tait (Eds.), Online and distance education for a connected world (pp, 293–319). UCL Press. https://doi.org/10.14324/111.9781800084797

Richardson, L. (2000). New writing practices in qualitative research. Sociology of Sport Journal, 17(1), 5-20. https://doi.org/10.1123/ssj.17.1.5

Stodd, J., Schatz, S., and Stead, G. (2023). Engines of engagement: A curious book about generative AI. Sea Salt Publishing. https://seasaltlearning.com/engines-of-engagement-generative-ai-book/

UNESCO (2019). The 2019 UNESCO Recommendation on Open Educational Resources (OER): supporting universal access to information through quality open learning materials. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000383205?posInSet=3&queryId=c4f022fb-4bef-4b66-ba53-3aafd4d586ed

UNESCO (2024). Draft Dubai Declaration on OER: Digital Public goods and emerging technologies for equitable and inclusive access to knowledge. UNESCO OER Dynamic Coalition Portal. https://oerdynamiccoalition.org/resources/draft-dubai-declaration-oer-digital-public-goods-and-emerging-technologies-equitable-and#draft-dubai-declaration-on-oer-digital-public-goods-and-emerging-technologies-for-equitable-and-inclusive-access-to-knowledge

United Nations. (2015). Transforming our world: The 2030 Agenda for Sustainable Development (A/RES/70/1). https://digitallibrary.un.org/record/3923923?v=pdf

Urmeneta, A., and Romero, M. (2024). Creative applications of AI in education. Palgrave Macmillan. https://doi.org/10.1007/978-3-031-55272-4

Appendix

1. Transparency                 

1.1 The crystal clear sea (Author 1). This shows the golden sand seen through tiny waves.	1.2 Transparency differs based on the intended purpose and design of the window (Author 2). This shows two arched stone church windows: one with clear glass and the other with multi-coloured, stained glass, and several people.	1.3 The Minaretein building in Doha (Author 3). This shows a stunning, white, futuristic university building with two huge beacons reaching into the blue sky.	1.4 Drinking Glass – clear and see-through, yet undeniably present and tangible (Author 4).

2. Translation

2.1 Providing fruits (Author 1). This shows a hanging pot plant with long, thin, strings of stems and leaves.	2.2 A globe provides the total view but will not help us navigate our immediate landscape (Author 2).   This shows a globe of the world placed on the ground in a street.	2.4 Writing doodles (Author 4). This shows a piece of lined A4 paper covered in rows of bright coloured dots and lines.

3. Teamwork

3.1 Fish out of water (Author 1). This shows an unusually shaped white fish, laying alone on sand.	3.2 An undisturbed puddle clearly reflecting the surroundings (Author 2). This shows a clear reflection of part of the London Eye in a puddle of water on the pavement.	3.4 Mirror – reflecting back what we project onto it (Author 4). This shows a round, black- edged mirror placed against a peach-colored background.