The Generative Fictions Exhibition: Words, Images and Territories (Sandra De Berduccy)

In June 2024 (winter in the Southern Hemisphere), on the busy outdoor promenade of the Centro Cultural Gabriela Mistral (GAM) library in downtown Santiago de Chile, images relating to the AIAI project were included in an exhibition called ‘Ficciones Generativas: Palabras, Imágenes, y Territorios’ [Generative Fictions: Words, Images, and Territories], convened by the Chilean research organisation Núcleo FAIR [Futures of Artificial Intelligence Research] and curated by Wolfgang Bongers and Valentina Montero.² This collective exhibition aimed to explore the sociocultural impacts of generative AI, and for our contribution, we wanted to privilege revealing the process of our experiments into the shortcomings of commercial image-generation models and our creation of our own image-generation tool, with which we could train models based on our own works and cosmovisions, over showcasing any resultant ‘works’.

We thus selected two moments in this collective research process. The first moment centred on the first stage of the project in mid-2023 where we generated images with Midjourney and each participant discussed their responses, in our focus groups, to the images generated with their prompts. Some of their responses were then edited by Sebastián Gerlic and made available on YouTube.³ This section of the exhibition aimed to enable the participants of the AIAI project to share their ideas and reflections on generative AI directly with the audience. We selected ten of the images discussed in the videos, printed them in large formats and accompanied them with a panel giving a brief biography of the person whose prompt had been used, an image of that person, the prompt itself and a QR code linking to the aforementioned video (see Fig. 18.1). In so doing, we were trying to be clear that these images were not being presented as artworks signed by their respective artists, but rather as experiments in AI image generation that had sparked reactions and debates about the possibilities and limitations of these tools among those involved in the project.

A second moment focused on the in-person workshop in Leeds in March 2024 with the Immersive Networks Collective where we collectively developed the IndigenIA prototype. We selected a series of photographs of the workshop in order to show the process of the prototype’s ‘generation’ as well as screenshots of the subsequent process of image generation with the prototype showing how we handled remote collaboration between Indigenous artists in Abya Yala/Latin America and members of Immersive Networks in the UK, in whose studio in Hebden Bridge the laptop hosting the IndigenIA prototype is currently located. This included some of the first images generated with our own models, based on textiles from our communities or on our own photographs, as we pursued self-expression through greater control and understanding of the technical process of image generation (see Fig. 18.2).

In terms of the success of our curatorial decisions, it is worth analysing the dynamic between the public, the materials and the exhibition space itself, which was an exterior walkway, a typical transit area rather than an environment designed for contemplation; and while the public was arguably composed mainly of visitors entering the cultural centre, importantly, the exterior location also opened the exhibition up to a wider public that would never normally set foot in such an elite institution. Furthermore, the strategic arrangement of the materials transformed this liminal space into a place of visual and conceptual encounter, where members of the public were engaged by the formal and thematic diversity of the exhibited pieces. Far from functioning as mere decorative elements, the materials generated active curiosity: passers-by were clearly initially attracted by the visual impact of the materials, but they were then seen to engage with the layers of meaning that transcended their immediate visuality. This voluntary pause demonstrated a participatory reception, reinforced by the use of QR codes that provided access to complementary audiovisual content. Thus, the proposal not only inhabited the physical space but also activated an intermedial experience, inviting the viewer to move from the act of seeing to that of inquiring.⁴

Science Week at the National Botanic Gardens of Ireland (Andreas Rauh)

A few months later, in November 2024, another set of materials from the AIAI project was exhibited in Ireland’s NBG. The exhibition, entitled ‘Artificial Intelligence, Art and Indigeneity: An Exhibition of Indigenous Art Made with Generative AI’, was part of Science Week Ireland and was set up in the NBG’s events gallery, a space usually reserved for temporary exhibitions that explore scientific and/or artistic themes. This setting intentionally placed the AIAI materials at the intersection of visual art, science and technology. The exhibition had three related objectives: to present the AIAI images in a context typically reserved for art and science, which thus fostered both contemplation and intellectual enquiry; to encourage reflection about the process of co-creation between humans and generative AI models; and to highlight Indigenous perspectives and encourage critical engagement with the practices of image-making and the influence of non-Indigenous elements (tools, aesthetics and technology) in shaping Indigenous visual representation.

The exhibition was divided into two parts: an introductory panel with a quiz about generative AI image-making processes and a main display that drew from the pool of images generated during the first stage of the AIAI project (see Fig. 18.3 top). The introductory quiz, originally designed for a ‘family fun day’ organised at the University of Leeds, asked visitors to distinguish AI-generated images from photographs. During this activity, facilitators explained how generative AI models work, introducing topics such as data scraping and intellectual property, the logic of diffusion models and the mechanics of prompting. After this introduction, audiences proceeded to the rest of the gallery space and the AIAI images.

The main part of the exhibition displayed twenty-five A2 posters in groups of five, on stands covered in large-scale photographs of the Botanic Gardens, a format that deliberately distracted from the visual impact of the AI images themselves (see Fig. 18.3 bottom). The layout encouraged free movement around the exhibition, including close examination of image details and broader views of sets of posters. The A2 print format was designed to encourage viewers to spend time with the images, in marked contrast to the fleeting consumption of mass-produced, low-resolution, AI-generated content available online (that is, AI slop). Each poster also integrated four key pieces of information: the image featured at the top and took up most of the area; below was the text prompt used in Midjourney to generate the image, the name of the prompt’s author and the version and date of Midjourney used (see Fig. 18.4). Overall, the poster layout with its integration of production data alongside the image intended to highlight the co-creative aspects of generative AI visual production and aimed to identify the contributions of each stakeholder (human and non-human) involved in the process.

The exhibition was designed to provide visitors with an understanding of methods and processes in generative AI image-making, thus leading to a reflexive and critical engagement with the AIAI works, and the arrangement in the room created opportunities for reflection on individual images and groups thereof, the production process, cultural and social contexts, and more. Furthermore, the exhibition aimed to draw from the reflexive qualities typically associated with art, as well as the intellectual enquiry associated with science, to encourage a thoughtful engagement with the AIAI images and the Indigenous involvement in their production. These curatorial decisions helped frame how the AIAI works could be approached while still leaving room for individual interpretation.

During its run, the exhibition attracted a broad audience, from regular NBG visitors to school groups participating in the Science Week programme. The timing of the exhibition also coincided with a moment of high public anxiety about the implications of generative AI.⁵ Younger visitors, particularly those familiar with generative AI tools, approached the quiz and images with less suspicion and more confidence. Older visitors tended to offer a more reserved set of critical reflections, as attested by the anonymous written feedback received. For instance, on the issue of creativity and co-production between humans and non-humans, many visitors identified the opportunities offered by generative AI models to prompt further imagination and discussion: ‘Humans feed into AI, which feeds back to humans, and prompts further human ingenuity etc’. However, responses also included: ‘I fear if AI becomes the norm our own individual creativity will be lost’, and, even though these might be ‘beautiful images’, ‘I miss the soul – the connection – between the viewer and the art’.

Underlying the written feedback and discussions with visitors was a broader sense of awe and anxiety, revealing a tension along a utopian-dystopian spectrum. Some visitors recognised the strong aesthetic and visual appeal of the images, alongside the promise of greater agency and autonomy in image generation. In the words of one self-identified fine artist, this was a ‘powerful exhibition. Very uplifting. The spirit in the pieces is soothing, magical, engaging, other worldly yet very human, provoking human feelings of great depth and meaning’. Others, however, were more ambivalent and highlighted the uncertainty and uncanniness raised by the images as ‘creepy yet beautiful. Eye opening. An interesting blend of words and images’. The most alarmist responses also raised issues beyond creative expression, such as, ‘A.I. kills humanity!’ and ‘causes harm to the environment. Please never use it in an exhibition again, instead commission real artists’.

A reflection offered by an Indigenous person who attended the exhibition helps contextualise these tensions further. They noted that dominant Western representations frequently portray Indigenous Peoples in either a mythical past or a utopian future (often an idealised life in harmony with nature), thus excluding them from the present.⁶ They noted this tendency was visible in the images themselves, as well as in the premise of imagining an Indigenous future with generative AI, which risks erasing contemporary Indigenous struggles and potentially reinforcing colonial narratives that emphasise the role of technology in social change. This is, of course, the logic of dominant discourses about AI, where rapid technological development is framed as inevitable and necessary to shape the future, thus echoing earlier colonial discourses (Hao 2025). In both cases, the future is imagined as predetermined by external technological forces rather than by collective social action. In the light of these dominant discourses, what we really need is greater critical reflection on new technologies, such as generative AI, as well as the articulation of alternative futures that address diverse needs and visions. This is what we hope the AIAI project and its different exhibitions have contributed to addressing.

Notes

1. Interesting research is just starting to emerge about the curation of AI art (see, for example, Pfefferkorn 2025), and Turkish artist Refik Anadol will open the world’s first AI art museum, dataland, in Los Angeles in spring 2026 (https://dataland.art/).

2. See Julio (2024) for an extensive review of the exhibition.

3. For the full playlist, see Mensagens da Terra/Messages from the Earth (2023–2024).

4. The inauguration of the exhibition also created an opportunity for local AIAI team members Azul and Loreto to attend and engage with the public directly.

5. In an Irish context, see the statement by An Roinn Fiontar, Turasóireachta agus Fostaíochta (Department of Enterprise, Tourism and Employment) (2024).

6. Pace the Indigenous visitor to the exhibition, we would refute the prevalence of dominant Western representations of Indigeneity featuring utopianist future scenarios, unless what they had in mind is something like the representation of Indigeneity in the Avatar film series (see Chapter 16). See also Tadeu’s comments on the subject of the prevalence of dominant Western representations of Indigeneity featuring the (mythical) past in Chapter 10.