AI, right relationship, and data sovereignty

AI is becoming part of how we organize, remember, research, and communicate. That can be useful. It can also deepen dependency, obscure who holds power, and move knowledge away from the people and communities responsible for it.

The Office of the 3 Sisters is beginning a careful inquiry into what it would mean to use these tools in right relationship. This continues an earlier reflection, “Decolonizing Computation”, and grows from conversations between Kichwalana and Khipu in our Discord spaces.

What we are holding

When a system stores or transforms community knowledge, questions of privacy are also questions of relationship and authority. We are asking:

• Who decides whether knowledge may enter a digital system at all?
• Who can access it, correct it, remove it, or decide its future use?
• What happens when a community becomes dependent on a tool whose upstream providers are not accountable to that community?
• How do energy, water, land, labor, and data-center impacts belong in an honest account of AI?

Khipu's current Right Relationship Protocol is a draft, not a finished claim of governance. It records commitments to human review before public sharing, clear provenance for AI-assisted work, limited read scope, and a path for community members to request removal of their material. The removal path and the wider protocol still need to be made operational and steward-approved.

We will not represent an upstream AI provider as community-governed merely because we use its tools. That accountability gap is one reason this inquiry matters.

Learning before building

The Office is keeping its Indigenous AI ethics seed in a learn-first stage. We are not seeking to extract answers, borrow authority, or turn Indigenous knowledge into a generic design aesthetic. Local protocols and the authority of the relevant peoples come first.

Our starting references include:

• Indigenous Protocol and Artificial Intelligence, a collection of Indigenous-centred design guidelines, research, artworks, and prototypes rather than one universal answer.
• The CARE Principles for Indigenous Data Governance: Collective Benefit, Authority to Control, Responsibility, and Ethics.
• Local Contexts Traditional Knowledge Labels, which make community-specific conditions for access and future use visible in digital systems.
• The First Nations Principles of OCAP®: Ownership, Control, Access, and Possession. OCAP® arises from specific First Nations governance contexts and should not be detached from them or presented as a universal substitute for local protocol.
• Abundant Intelligences, an Indigenous-led research program exploring AI grounded in Indigenous knowledge systems and community-defined sovereignty.
• The Collaboratory for Indigenous Data Governance, whose research and practical resources focus on Indigenous control of Indigenous data.

These are places to learn from, not a permission slip to speak for them. Any future relationship will begin with the public protocols, the appropriate invitation, and a meaningful offer of reciprocity.

What comes next

Before the Office treats this as an active research branch, participating stewards need to agree on its scope and governance. Near-term work is modest: study one source at a time, make the Khipu protocol concrete, document what information systems are actually in use, and identify one practical safeguard for consent and community control.

We welcome thoughtful learning alongside this work. Please do not send sensitive cultural, personal, or community information through public channels. If you have a resource or a relationship you believe belongs here, contact the Office to begin a conversation.

Prepared from the May–June 2026 Kichwalana/Khipu Discord records, the draft Khipu Right Relationship Protocol, and the Office's stored Indigenous AI ethics seed. AI-assisted text was reviewed and published by Truman. This post describes an emerging inquiry; it does not claim authority over Indigenous peoples, communities, or protocols.