Small AI: 3 Things to Explore

Inspired by the counter-movements of slow fashion and slow food, Slow AI project investigates three emerging AI counter-narratives – Small AI , Ancestral AI & Esoteric AI  – and explores what it might look like to incorporate them into our everyday practice.

Small AI questions the dominance of large-scale models by addressing environmental impact, discriminatory language, and cultural preservation. It’s an invitation to use notions of fractals, friction, and fragmented-ness in nature and math to inspire ideas around Small AI and send ripples that shift practices around Big AI. It advocates for equitable, sustainable alternatives, prompting exploration of community-focused governance, and interconnectedness as pathways to more just futures.

In this update, we're excited to share three insightful elements to deepen your understanding of Ancestral AI. First, explore a recent reading that captivated our research lead, Nadia Nadesan. Second, meet a person of interest making waves in the field. Lastly, watch a compelling snippet to further immerse yourself in the research of the Small AI project. Dive in and explore these curated highlights from Nadia.


Research Lead


Jun 2024


Stimuleringsfonds Creative Industries + currently looking for additional co-funding

decroative corner decroative corner
decroative corner decroative corner


The Te Hiku Media Corporation’s LLM has been an inspiration for Small AI, especially in terms of how situated environments and cultures enable more life-affirming technologies. Nan O’Sullivan’s article highlights working towards integrating Indigenous knowledge in design education and moving towards a more ‘place-based’ pedagogy informed by traditions that prioritize the health and well-being of people and the place around them. O’Sullivan provides a history of bringing in Maori culture and heritage as a part of design education in New Zealand not only as a precolonial practice but as a modern mode of creating and sharing design practice and knowledge. In designing especially with technology, the bottom line is to make a profit by designing around those who have wealth. However, students in pursuit of developing their own perspective and design practice were asked to acknowledge ‘kaitiataknga’ (guardianship), and manaakitanga (respect and reciprocity) when engaging with real-world issues. Te Hiku Media and the Maori community around it have created modes of working and technology that demonstrate that it is possible to develop technology for and with people. The article by Nan O’Sullivan further elucidates that integrating Maori knowledge and learning is part and parcel of sowing the seeds for creating tech that creates the possibility for multiple narratives and a plurality in ways of being.

decroative corner decroative corner
decroative corner decroative corner


We don’t go and investigate the killing of somebody, because we find it interesting. We only never want to enter and break into a traumatic space of bereaved family uninvited. So we only work if the family asks us for it. ‘

Eyal Weizman is the founder and director of Forensic Architecture. He also is a professor of Spatial and Visual Cultures at Goldsmiths. On 9 May 2024, Weizman came to Elisava’s Design School in Barcelona to deliver a talk on the work of Forensic Architecture. During his talk, Weizman addressed how cartography was a tool of the colonizer. Cartography or drawing maps and annotating the terrain is to create a graphic representation of the violence that has transpired or to be transferred onto the land. In his words, ‘There is no violence without the representation of violence, the representation of violence and violence are completely entangled.’ Cartography forms an integral part of Weizman’s work as an architect. The reason that Weizman’s work appealed to me was how he chose to reframe architecture as a means to support vulnerable communities. Forensic Architecture reorients architecture as a method to recreate scenes to investigate violence such as the killing of journalist, Shireen Abu Akleh, covered in Weizman’s talk. Forensic Architecture has been working with Palestinian and more recently Namibian peoples. Within their work, he clearly stated that they do not do the investigations because they find it interesting. After all, they do not want to ‘break into’ the traumatic space of a bereaved family uninvited. So they only work when asked by the community or family. In the case of Shireen Abu Akleh, they were able to recreate the scene and even find the extra bullets that were aimed at her and shot into the tree she was standing next to. However, rather than extracting the bullets they let the tree live and remain there until justice becomes possible for her. Forensic Architecture orients architectural methods and tools like AI to affirm the existence of communities. They work with people, their families, and their grief. I found Forensic Architecture to be a point of inspiration for Small AI as a reference for small-scale relational dynamics and practices that create impactful and inspiring narratives and knowledge that challenge today’s injustices.

‘And we all agreed, we’re not going to cut down the street to retrieve the bullets. That tree is where they are. And this is going to be preserved … Number two, this is where she died and where the bullet is going to remain unfilled justice is going to become possible. And sometimes the life of a tree is also more important than the life or the action that you can make with evidence with it. And I think that it is extremely important to know also, that in situations like those who work in Gaza every day, the mechanism, the forums of international justice do not exist. We are recording things [for the future]‘


I came across the Potato Internet by Caroline Sinders through Internet Teapot’s Zine Making Session on Small AI. Caroline Sinders is a machine learning design researcher. In this video, she explains her work in collaboration with Trammel Hudson to create an internet powered by potatoes. Through the work Sinders and Tramell explore the idea of small ways to build different kinds of technology and hardware. She argues that alternatives to large scale platforms like Twitter and Instagram may not be one-to-one. In that there may not be a large scale alternative that functions like these large platforms, but instead an array of smaller ones. Sinders through her work then presents in this work how slow and friction might be integrated in the function and design of the technology itself. One of the ideas to come from this work is of a feminist data set and reframing data as organic and material along with the idea of using recycled material or what would have otherwise been e-waste. How might the development of new technologies be informed by what works for our ecosystems and reusing and remaking from devices, machines, and materials from the past?

Caroline Sinders (in collaboration with Trammell Hudson) - The Potato Internet (2022) from werkleitz on Vimeo.