We are busy working on a position paper on Indigenous AI, drawing on the conversations that happened in the Indigenous Protocol and AI Workshops. We are collaborating with our main funder, CIFAR, to make this document public in the spring of 2020.
“Aloha is the intelligence with which we meet life.”- Olana Kaipo Ai
The position paper on Indigenous Protocol and Artificial Intelligence (IP AI) is a starting place for those who want to design and create AI from an ethical position that centers Indigenous concerns. It is an attempt to capture multiple layers of a discussion that happened over 20 months, across 20 timezones, , during two workshops, and between Indigenous people (and a few non-Indigenous folks) from diverse communities in Aotearoa, Australia, North America, and the Pacific. Our aim is to articulate a multiplicity of Indigenous knowledge systems and technological practices that can and should be brought to bear on the ‘question of AI.’ To that end, rather than being a single unified statement this position paper is a collection of heterogeneous texts that range from design guidelines to scholarly essays to artworks to descriptions of technology prototypes to poetry.
The Position Paper contains the following:
The Contexts section speaks to the intellectual and cultural currents running throughout the workshops.
Workshop Description IP AI Working Group
Information on who was involved as organizers and participants, the goals set out for the workshop series as a whole and the two separate workshops specifically, the agenda for each workshop, the main funders and supporters, and a brief summary of the events that led to the founding of the workshop series.
AI: a new (r)evolution or the new colonizer for Indigenous peoples? Dr. Hēmi Whaanga
This is an essay by linguist and te reo Māori specialist Dr. Hēmi Whaanga (Ngāti Kahungunu, Ngāi Tahu, Ngāti Mamoe, Waitaha). Dr. Whaanga warns of the potential for AI and related technologies to be used against Indigenous peoples as an extension of colonial practices of exploitation, extraction and control, particularly those that displace a peoples’ understanding of themselves with a worldview that favors the colonizer. He discusses issues of data sovereignty in a technological landscape populated by AI systems existentially dependent on sucking up vast amounts of data on human activity, thereby putting Indigenous traditional knowledge and customary practices at risk of global-scale appropriation. Dr. Whaanga finishes his essay with a call to centralize Indigenous concerns in the work of establishing global ethical guidelines for the design and deployment of AI.
The IP AI Workshops as Future Imaginary Jason Edward Lewis
This essay positions the workshops as a rich example of how to collaboratively create shared future imaginaries. Drawing on internal notes taken throughout the first workshop, Lewis reviews the breadth of professional and cultural backgrounds, the many different types of conversations, and the concerns as well as hopes of the participants to paint a picture of the multiple layers of complex knowledge exchange that took place. He also articulates a number of different ways that Indigenous knowledge already is reflected in technological practice and the visions participants shared about how their particular community’s cultural practices could provide frameworks for designing aspects of AI systems.
The Vignettes section gathers together five different visions of how AI might be built according to values articulated in Anishaabe, Coquille, Kanaka Maoli/Blackfoot, and Euskadun epistemologies, respectively.
Gwiizens, the Old Lady and the Octopus Bag Device Scott Benesiinaabandan
This short narrative is an AI ‘creation’ story that illustrates how new technology such as AI might be incorporated into and made of a piece with the existing canon of Anishinaabe creation stories, in a manner that makes use of existing methods for sharing knowledge while keeping it culturally grounded.
Gifts of Dentalium and Fire: Entwining Trust and Care with AI Ashley Cordes
This text argues for Indigenous people to seriously consider the use of blockchain combined with AI to help them manage their communities’ business, making the case that such technologies can be used to increase Indigenous sovereignty and self-determination vis-à-vis the hegemon. A member of the Coquille tribe, Cordes uses that community’s notions of ‘trust and care’ to ground her vision of how the technologies should be properly designed and to map out how they might be implemented. She also explores what it means to take seriously the admonishing to consider AI as non-human kin, including thinking about what the AI’s needs might be.
QUARTET Jason Edward Lewis
Lewis contributes “Quartet,” composed of a poem sequence and a short description illustrating how epistemological diversity within AI design might look. The texts imagine a future where young Kānaka Maoli are raised along with three AIs, each built according to different conceptual frameworks. One AI takes inspiration from Kanaka notions of land, responsibility, and family; another from the Blackfoot language’s basis in flow rather than objects; and the third from suppositions about how the octopus’s nervous system is organized to accommodate the semi-autonomy of its arms. The three AIs and the human work collaboratively to make decisions in support of Kanaka flourishing that take the environment, human and non-human relations, and past-present-future into consideration.
How to Build Anything Ethically Suzanne Kite
Kites’ contribution draws on Lakota knowledge frameworks to propose a protocol for ethically building computer hardware from the ground up. She discusses what it means to operate in the world in a ‘Good Way’ according to Lakota principles, and draws on how Lakota form relationships with stones to explore how we might form relationships with AI hardware (which one can think of as being made out of stones). She then maps out a process for building physical computing devices in a good way, using the protocol steps for building a sweat lodge as a guide. Kite closes with a sequential list of questions that should be asked at each step of creating such devices—questions that are designed to keep the building process aligned with the good way.
Wriggling Through Muddy Waters: Revitalizing Euskadunak Practices with AI Systems Michelle Lee Brown
Brown describes a Txitxardin Lamia, a biotech eel-AI developed from principles based on Euskaldunak (Basque people)-eel relations. She outlines a VR environment in which an elder Txitxardin Lamia would reside, where students could learn protocols for interacting with this elder and receive teachings that reorient them to more reciprocal ways of relating to, and being with, the world around them. Along the way she provides context, including a discussion of the long relationship between her people and these eels, the central role this relationship plays in coastal Basque culture, and the need to think through the materialities out of which we are creating and housing our AI systems.
The Prototypes section details the effort to implement an AI component technology using Indigenous values.
Indigenous Protocols and Artificial Intelligence in Action Caroline Running Wolf and Dr. Noelani Arista
This introduces the objectives of the project. Running Wolf and Arista identify the Indigenous values that the team shared, including respect, reciprocity and relationality. They also discuss how all team members have a commitment to Indigenous language revitalization. These commitments informed the team’s brainstorming about what kind of project they should and could undertake in the form of a ‘hackathon’ over the five days of Workshop 2. The brainstorming lead them to imagine an app that would recognize objects and provide the ‘ōlelo Hawai‘i words or phrases to describe those objects. Running Wolf and Arista then review the many different complexities of the development process itself, from the integration of local customary knowledge holders and language keepers from the beginning, to the challenges of using existing software modules that are based on non-Indigenous models of language-use, to the reliance on digital dictionaries that are often the flawed result of colonial historical processes, to the opportunities created by networked knowledge-sharing to shore up and validate language choices.
Indigenizing AI: The Overlooked Importance of Orality in Print Dr. Noelani Arista
This essay provides a deep and wide Hawaiian-rooted context for the Hua Ki‘i project while also articulating a conceptual model for Indigenous technological development that could be applied in other contexts. At its core, Dr. Arista’s contribution argues for the importance of aligning cultural and computational competencies so that each reinforces the other. She discusses how cultural competency might be best understood, describing how the interchange between Hawaiian customary knowledge, orality and print technologies mutually reinforce one another, and imagining how that dynamic might be extended to include computational technologies such as those used for language acquisition and translation. She observes how “we are entering a new phase of language revitalization where technology can assist Indigenous people in organizing data in ways that allow us to synthesize ancestral knowledge and rebuild systems of knowledge keeping and transmission”; the key will be to ensure that the intellectual architecture preserved orally and textually by our ancestors helps shape the computational architecture of our digital technologies—and the data on which they are fed.
Development Process for Hua Ki‘i and Next Steps Michael Running Wolf, Caroline Running Wolf, Caleb Moses and Joel Davison
This section includes details such as the design of the user interface, the app architecture, and envisioned usage, as well as the software modules used for language and image processing. The authors also discuss how the prototype sets the stage for further development.
Dreams of Kuano’o and the Road to Kuano Michael Running Wolf
“Dreams of Kuano‘o” is a short story that imagines a future where Hawai‘i has regained its sovereignty, and where the ‘Queendom’ requires all visitors to use an AI app called Kuano‘o to guide them while visiting. It touches on how sovereignty might be enforced and sustained using such an app, including compulsory education in the island’s history and cultural norms; penalizing the use of languages other than ‘ōlelo Hawai‘i in most public contexts; and use of social credit scores to modulate behaviour. The second component discusses the challenges of moving from the Hua Ki‘i prototype made by the team to the Kuano‘o AI system envisioned in the short story. Foremost among these are obtaining and maintaining clean data to use in the training of the necessary speech and image recognition systems.
Appendix A: Pre-Workshop Blog Posts & Workshop Interviews
Appendix B: Reading List
Appendix C: Biographies
Appendix D: Workshop Schedules
Painting by Sergio Garzon. IP-AI © 2019.