Thea Pitman

In a similar manner to efforts made by people working in the ‘AI for good’ sector to determine ways that AI technologies might help Native American, First Nations, Inuit, Aboriginal and/or Māori communities, early interest by groups such as UNESCO in the implications of AI for Indigenous Peoples in Latin America and the Caribbean has tended to focus first on ethical issues of increased inequalities. This is because of unequal access and literacy in relation to the technologies involved, the dangers of algorithmic bias (encompassing bias in the algorithms themselves as well as the datasets used) and the impact on Indigenous data sovereignty where AI tools are used for enhanced extractivist practices by the Global North. They have then gone on to identify a range of practical applications, such as the use of AI for Indigenous language revitalisation, for documentation and preservation of cultural heritage and, somewhat ironically given the already well-known environmental cost of AI technologies (cf. Crawford, 2021), for dealing with global challenges such as environmental stewardship. See, for example, the now disbanded IVOW (Intelligent Voices of Wisdom) AI’s ‘interactive whitepaper’ entitled Bringing Cultural Context to Artificial Intelligence (2020) and Luz Elena González Zepeda and Cristina Elena Martínez Pinto’s UNESCO report, Inteligencia Artificial centrada en los Pueblos Indígenas: Perspectivas desde América Latina y el Caribe [Artificial Intelligence and Indigenous Peoples: Perspectives from Latin America and the Caribbean] (2023), as prime examples of this trend.¹ None of these reports, however, cover the period since generative AI tools were made publicly available.

More recent coverage of this topic by Indigenous radio producer Anápuàka Tupinambá Hãhãhãe generally takes a similar line, updated to encompass the applications of generative AI. In one article (Tupinambá Hãhãhãe 2023), the author starts by outlining questions of unequal access, the danger of loss of control and/or sovereignty over Indigenous knowledge because of data extractivism and cultural appropriation, increased stereotyping (because of algorithmic bias), as well as generative AI-fuelled loss of jobs and income streams (particularly for Indigenous artisans and artists). He then goes on to argue that AI can be of use in the preservation of Indigenous cultures, in improved access to education and healthcare, and in supporting sustainable development and entrepreneurial initiatives – just so long as Indigenous communities are involved in decision-making, development and implementation of AI tools and projects, in designing ethical guidelines for the use of AI in their communities, and are provided with training programmes in the use of AI tools. In a later piece, Tupinambá Hãhãhãe (2025) considers the value of open-source generative AI tools and the development of bespoke tools for Indigenous people, especially creatives, explaining how he personally uses AI tools ‘como uma ferramenta accesível para ditar textos por voz, corregir e formatar conteúdos, além de gerar imagens para ilustrar informações. Isso me permite superar barreiras da dislexia e potencializar a comunicação indígena com autonomia e inovação’ [as an accessible tool for dictating texts, correcting and formatting content, as well as generating images to illustrate texts. This allows me to overcome the barriers of dyslexia and enhance Indigenous communication in autonomous and innovative ways]. This ‘illustrative’ mode that facilitates communication is one of the notable trends in Indigenous uptake of generative AI.²

In terms of the ways in which Indigenous artists in Abya Yala/Latin America have engaged with (particularly generative) AI tools, the pattern generally follows that outlined in Chapter 3 in relation to Indigenous artists in the Anglophone settler (post)colonial world, spanning revisionist and speculative, futurist approaches. The earliest work started appearing in 2022, very much riding the wave of the hype surrounding the public launches of DALL-E and Midjourney, and, perhaps not unsurprisingly in terms of early access to these technologies, it has tended to be led by artists active on the international art scene who originate from Indigenous communities in the region. These artists include Purhépecha-Italian performance artist Tiara Roxanne (currently based in Germany) who seeks to explore the ‘ethics of AI through a decolonial cyberfeminist lens’ as she says on her artist’s website (‘About’, Roxanne 2025) and whose work looks at questions of computer vision, surveillance and data colonialism (for example, ‘My Face Is Not A Barcode’ [2022]); and Guaraní multidisciplinary, ‘AI-collaborative’ artist Kira Xonorika (Paraguay; now based in USA) who creates AI images of surrealistic landscapes with colourful, often amorphous figures that seek to challenge binary discourses around gender, sexuality and cultural identity (Murphy 2023). Xonorika also organised the Swiss Arts Council-funded Future Memory Lab art residency in Asunción in 2024, which claimed to ‘be a first in South America’ focusing on AI and exploring questions such as,

How can analog traditions inform and enrich contemporary practices in AI and generative technology? In what ways can AI contribute to the preservation and revitalization of traditional cultural and artistic practices? How can intersectionality and decolonial epistemologies shape the development and application of generative technologies? What role does AI play in imagining alternative futures and constructing possible worlds? How can generative arts and AI expand our understanding of autonomy, agency, and creativity? (‘Future Memory Lab’ 2024, ‘Artistic Residency’)

A final, more localised example of the speculative, futurist mode in AI art is the work of K’ita Wichhu (artist’s name of Noemi Gonzales Cabrera), a mestiza artist from El Alto, Bolivia, currently engaged in reconnecting with her Indigenous heritage. K’ita Wichhu has used AI image-generation tools to help imagine futuristic scenarios for a ‘visual performance essay’ called Post-humanas y sueños cyborg de + seres en los Andes [Post-human Women and Cyborg Dreams of + Beings in the Andes] (2023) that she then staged and photographed in collaboration with a group of Indigenous women in the Altiplano. In this example of what she calls ‘Andean futurism’, AI serves as ‘un puente entre la imaginación y lo realizable’ [a bridge between imagination and reality] and a tool that helps her ‘soñar un futuro suspendido’ [dream a potential future] (Cisneros 2025).³

Non-Indigenous international artists have also recently sought to involve Indigenous communities in generative AI art projects. For example, in 2023, Turkish media artist Refik Anadol, working through the Instituto Nixiwaka, visited two Yawanowá communities in the western Brazilian Amazon who collaborated with him to create a series of AI artworks combining environmental data and drawings made by Indigenous community members that were then transformed into a mesmerising AI animation reminiscent of the hallucinatory effects of the popular early AI neural network DeepDream. Subsequently, 1,000 digital stills were minted as NFTs⁴ and sold to raise $1.5 million, the proceeds of which went to the communities (Yawanowá and Anadol 2023). While in interview Cacique Nixiwaka Yawanowá was very positive about the impact of the project on his community (Scorpios Encounters 2023), for a very critical review of this project see R. H. Lossin (2025) who writes,

Rather than the adage ‘garbage in, garbage out,’ favoured by computer scientists and statisticians, AI’s transformation of inputs into visual objects is a matter of ‘value in, garbage out.’ Art collection in, garbage out; literature in, garbage out; apples in, garbage out; human subject in, garbage out; Indigenous lifeways in, garbage out. (5)

And Lossin concludes, ‘Given the potential material impact of these projects, Anadol’s repeated promises to produce some future benefit are stupendous evasions of responsibility to the present … typical of a techno-politicial mindset’ (5).

The revisionist mode has also emerged as a compelling use of generative AI tools for the representation of marginalised communities in the region thus attempting to creatively fill gaps in the history of visual culture – portraits that were never painted, photographs that were never taken, precisely because of the marginalised social status of the subjects.⁵ In an Indigenous context, Mexican artist Rogelio Séptimo’s Exhumar la memor.IA [Exhuming memory] (2023) participatory art project combines drawing, testimonies and analogue photography with generative AI to create a historical record that never existed of his P’urhépecha community on the island of Janitzio in Lake Pátzucaro (Prada 2025, 3273–74). Additionally, Quechua filmmaker and game designer Violeta Ayala (Bolivia, now based in Australia) has produced a very striking series of AI-generated photographs with DALL-E called Las Awichas [The Grandmothers, in Quechua] (2022), intended to playfully invent the non-existent archive of images of older Quechua and Aymara women.⁶ This was premiered at the Bolivia International Digital Art Fair at mARTadero, Cochabamba, in September 2022, with all signage in Quechua and Spanish. It has since been developed into an augmented reality game on 8th Wall and was presented as part of the ‘GLoW: Illuminating Innovation’ exhibition hosted by King’s College London along The Strand in 2024.⁷

While some of the AIAI team members became aware of Violeta Ayala’s work during the first stages of the AIAI project, as both Las Awichas and work from other projects that Thea or Thydêwá had sponsored (ORIGEM and Mujeres invisibles, respectively) were presented at the Bolivia International Digital Art Fair, and aruma knew Violeta personally as they had attended school together in Cochabamba before the latter emigrated to Australia, the artists, writers and others involved in the AIAI project were almost unanimously unaware of other examples of the way artists, Indigenous or otherwise, were starting to use these tools. However, a real forerunner in the field of portraiture with AI art tools, and a very direct antecedent of the AIAI project, was Sueño profundo [Deep Dream] (2021), part of the AIRE project for virtual co-creation between Indigenous and non-Indigenous artists organised by Thydêwá during the period of lengthy COVID-19 lockdowns. The two-minute animation, edited by aruma in collaboration with artist and curator Naine Terena and with music by Mocoví singer Shiraigo Lanche, animates a series of selfies taken by the artists involved (Anna Campagnac, Ângela Berlinde, aruma, Carina Dessana, Sheilla Souza, Shiraigo Lanche, Valeria Scornaienchi and Naine Terena) by exposing them to the programme DeepDream.⁸ The animation was one of the most striking outputs of the project, and the consonance between dreaming as it refers to this kind of generation of images based on pattern recognition and to Indigenous desires to dream as a way of connecting with the spirit world, as well as to imagine future scenarios that include their cultures, is compelling. It is this potential of generative AI tools to collaborate in processes of dreaming that makes them attractive for use in Indigenous contexts.