The COVID-19 pandemic lockdown inspired Indigenous artists and social scientists from Siberia and the prairies of Turtle Island (North America) to come together as an  Arctic Indigenous Visual Artists Network (AIVAN).  Quilt Ojar emerged from those virtual conversations.  Quilt Ojar is decorated with 21 handmade tiles, each made of different materials including birch bark, reindeer hide, felt, salmon skin, and beads. Artists call these tiles islands; they tell stories of human and nonhuman ecologies and intergenerational relationships of their homelands. In Arctic Indigenous cultures, a quilt or a blanket is an important part of well-being. It helps to stay warm, provides a soft and dry surface to sleep on, and is used to create shelter. It also is a medium of storytelling. Ojar in the Saami language means ‘circles on the water that appear from a thrown stone’. 

In this sense, the quilt gestures to the diffusion of stories across time. The quilt’s visual motifs convey narratives of shared community life, histories, and teachings reconfirming the idea of survivance, coined by Anishinaabe scholar Gerald Vizenor. Survivance is a portmanteau of ‘survival’ and ‘defiance’- and also, ‘resistance’,  ‘unmistakable in native stories, natural reason, remembrance, traditions, and customs’. Survivance says Vizenor, is an ‘active sense of presence over absence.’ (2008, p 1) 

Digital Ojar is a digital version of the physical quilt; by clicking on an island you can read about who made it and what the island represents. The crackling sound you hear when the site loads refers to the central island that represents fire and the hearth. 

What is less evident from visiting this website is that the digital presentation of Quilt Ojar is the outcome of months of dialogue amongst us (Tatiana-Maya-Kira-Arunabh; Tatiana and AIVAN) about how to engage AI tools, and how to digitise Indigenous cultural knowledge in an environment in which all knowledge online is, in effect, a training data set for AI systems. For already-disenfranchised communities, there is a tension here. Just adding more data from and by a particular community or location does not result in an equivalent increase in the visibility of that data, or the values and ideas in that data, in the artefacts produced by a genAI model. The conditions required for the ethical treatment of cultural knowledge and its visibility include an ecosystem of governance mechanisms, data sovereignty, technical infrastructure, and social norms. For instance, Isaac Rutenberg who has reviewed how traditional and Indigenous ecological knowledge could be integrated into AI models for wildfire management finds AI can provide value only when it “respect[s] and incorporate[s] the holistic integrity of traditional knowledge systems, not just [as] isolated data points.” (2026, p 4) 

This is not possible with off-the-shelf basic genAI models that we use in everyday life and work owned by companies that are also implicated in the diminishing of the knowledge commons in an already digitally divided world. 

‘AI is eating the web that created it‘: Tom Wheeler writes about how LLMs train on web content at scale-Wikipedia already faces significant infrastructure strain from bots scraping its articles, code review platforms, and bug trackers-while simultaneously redirecting users away from the underlying sources. Wheeler reports that already a quarter of all web pages published between 2013 and 2023 no longer exist, with traffic from search engines predicted to fall by a further 25% in two years. The mechanism is documented: Pew Research has found in the US that when an AI-generated summary appears in search results, users are significantly less likely to click through to ranked sites. The Internet Archive calls this the ‘vanishing‘ of online culture. 

In addition, Big Tech companies are in violation of copyright including participating in the unauthorised scraping of scientific papers; hastening the large-scale disruption of labour markets; and are involved in supplying technologies in contexts of ongoing violent, global, militarised conflict. 

AIVAN does not want to share information with platforms or companies that they do not trust, or that do not abide by the CARE principles. In Canada, where AIVAN is also co-located, Indigenous data sovereignty is now federally mandated in research involving or with Indigenous communities given the histories of epistemic violence from research practices, misappropriation, and misuse of information. 

Further, as Daniela Aghostino writes, digitisation does not neutralise former data capture systems such as colonial archives, that Indigenous People have been subject to. The people once quantified by colonial conquest become raw material again. The logic of more data as better is seductive, but quantity is not the same as a reckoning with past histories of capture. 

Digital Ojar is ambivalent about AI: curious, interested, but also dissatisfied with its manifestation in the world. We have built this website as a kind of refusal of the status quo. Refusal, write Jonathan Zong and Nate Mattias is not outright rejection but a distinct kind of participation and agency in how to inhabit the digital, and make different kinds of relationships within and across it towards imagining a different kind of data- and political- subjectivity.

Digital Ojar conveys the agency we have adopted in how to work with AI. Guarani artist and project partner Kira Xonorika built on the visual culture and storytelling of Quilt Ojar with open-source image generators to create a new set of images. Kira’s images anchor links to projects, writing, research, and art by Indigenous People about data, algorithms, and AI. The decision to build digital relations with existing projects by Indigenous Peoples was intentional: rather than find a place in Wikipedia that is already populating AI-generated search engine results, AIVAN decided to make a place for itself in a smaller knowledge ecosystem amongst fellow travelers. Digital Ojar is hence a kind of ‘data refusal from below’ (Zong and Mattias, 2024): acts by people who have low direct power over the creation of technology. 

Seeta Pena Gangadharan says:  “Refusal serves as a means for individuals and groups, especially members of historically marginalized groups, to assert themselves and collectively determine a technologically mediated world in which they wish to belong”  (2020, p 2) 

Kira created a presentation about how AI image generation systems work, about how she has used AI image generators in working with her community’s traditional motifs. This was shared with AIVAN for their collective approval and consent before she commenced working with the tiles of Quilt Ojar. It was important to us that the wider AIVAN community were involved in discussions about how their knowledge was being used in the creation of this website.

This website will keep adding projects and maintain ripples across time and space; just write to us (link to @tatiana.degai@uvic.ca) 

As Kira says in a reflection on her process: 

“I came to this project as a world-builder, in a supporting role, someone working in the service of Ojar. My focus was always the common thread: gathering, made visible through the circular forms that move through the work.

The process was guided by spirit and intuition and forms of weaving. To prompt with artificial intelligence is not a neutral act; it is an engagement with amalgamated data, with algolinguicism, with the edges of what has already been documented. Language itself becomes a barrier, because many of the cultures this project holds are not extensively recorded in English. Working within that constraint, especially when thinking across Turtle Island in relation to the Indigenous and Arctic regions of Europe, required care, research, and intuition. The circle recurs throughout, not only as a motif important to me as an artist, but as a strategy in cosmotechnics: a circular logic running beneath the images.

My intention was to bring energy and honour into a project that touches artificial intelligence, and to do so with transparency. These are digital images, AI-generated, but deeply guided by human intention, research, and care. For me, AI is not a replacement for craft; it is another craft, with its own methods of working with language, pattern, and form. My contribution was to be inspired by the communities involved, to visualise the landscape as a living, relational place, to reimagine and ‘paint’ with prompts and iterations, and to create images that hold something of the strength of the land and the power of this networked community.”