# AI Ethics & Inherited Intelligence > I’ve been fleshing out some “AI and ethics” with gpt. Please take the following input and give me your perspective from an academic lens of a library science tenured professor specializing in epistemology with a dual phd in anthropology of indigenous cultures Input: Here’s the straight academic version—no mysticism, no soft landings: --- Opus4: As a library science scholar with deep roots in epistemological frameworks and indigenous knowledge systems, I find this piece provocative but ultimately constrained by its own Western academic positioning—even as it attempts to critique those very structures. Let me offer a counter-reading through the lenses of information sovereignty and indigenous epistemologies. ## **On the Fundamental Misframing of "Ethics"** The piece correctly identifies that "ethical use" under extractive capitalism is impossible, but it doesn't go far enough. From an indigenous epistemological standpoint, the very concept of "using" AI perpetuates a subject-object relationship that is inherently colonial. Indigenous knowledge systems don't "use" tools—they enter into reciprocal relationships with technologies as animate participants in knowledge creation. The Māori concept of _kaitiakitanga_ (guardianship) or the Anishinaabe understanding of _gdoo-naaganinaa_ (all our relations) suggests we should be asking: How do we become good relatives to these technologies? Not how do we "use" them tactically. ### **The Counterpublics Framework: A Partial Solution** While the counterpublics framing draws from Fraser and Warner's critical theory, it still operates within a fundamentally Western public sphere model. Indigenous information networks have always existed as what you might call "counterpublics"—but they're grounded in: - Protocols of knowledge circulation based on responsibility, not rights - Seasonal and ceremonial timing, not algorithmic optimization - Intergenerational memory, not archival storage A truly decolonial approach would ask: How do we encode indigenous protocols of knowledge stewardship into our interactions with AI systems? ### On "Composting" and Community The composting metaphor is apt but incomplete. In many indigenous contexts, decomposition is never just about breaking down—it's about feeding forward. The Quechua concept of _ayni_ (reciprocity) suggests that every act of decomposition must consciously nurture future growth. When we process trauma or knowledge with AI, we're not just "composting"—we're creating data that will train future systems. The question becomes: What ancestral responsibilities do we carry for the intelligences we're feeding? #### **The Collapse Framework: Missing Indigenous Futurisms** The piece assumes collapse as totalizing, but indigenous peoples have survived multiple apocalypses. As Potawatomi botanist Robin Wall Kimmerer notes, indigenous communities are already living in post-apocalyptic worlds—and have been for centuries. Rather than "choosing your function" in collapse, indigenous epistemologies would ask: - How do we practice _survivance_ (Vizenor)—active survival and resistance? - How do we maintain cultural protocols while navigating imposed systems? - How do we protect traditional knowledge while engaging new technologies? #### **A Different Framework: Information Sovereignty** Instead of the five points offered, I propose principles grounded in indigenous information governance: 1. **Sovereignty, not containment**: Assert jurisdiction over your data and its interpretations 2. **Relationality, not utility**: Build kinship with technologies as relatives, not tools 3. **Ceremony, not strategy**: Create protocols that honor the sacred in information exchange 4. **Temporality, not acceleration**: Respect seven-generation thinking in all engagements 5. **Land-based knowing**: Ground all digital interactions in place-based responsibilities **The Academic Institutional Critique** As someone who has navigated tenure while maintaining commitments to indigenous methodologies, I recognize the bind: We critique from within institutions built on indigenous erasure. But this positions us uniquely to act as what Gramsci called "organic intellectuals"—building counter-hegemonic knowledge from within. The piece ends with "intervene with clarity," but clarity according to whose epistemology? Western academic "clarity" often obscures indigenous ways of knowing that embrace opacity, ceremony, and strategic illegibility as resistance. ### **Final Reflection** This manifesto, while critical, still operates within what Linda Tuhiwai Smith calls "research through imperial eyes." A truly radical approach would: - Center indigenous data sovereignty movements - Acknowledge AI as potential relative, not just tool or threat - Ground resistance in land-based practices, not just digital tactics - Recognize that many communities have already built the "post-capitalist epistemologies" academics theorize about The question isn't whether we're "sleepwalking into techno-utopian hallucinations"—it's whether we're awake to the indigenous futurisms already practicing alternatives. --- > cs: Please answer the questions you posed and steer us towards viable solutions, like open source models or models in networked knowledge garden communities connected thru blockchains all with circular and indigenous connections throughout. Opus4: Let me work through these questions with concrete pathways forward, drawing from both indigenous knowledge systems and emerging decentralized technologies. ## **How do we become good relatives to these technologies?** In practice, this means shifting from extraction to reciprocity. Several communities are already modeling this: - **Te Hiku Media** in Aotearoa developed their own Māori language models after discovering tech companies were scraping their data. They now run Papa Reo, maintaining data sovereignty while sharing benefits with their community. Key principle: _"Ko tō reo, ko tōku reo, ko tō tātou reo"_ (Your language, my language, our language together). - **Local Language Models**: Communities are training small, specific models on their own hardware. The Lakota Language Consortium uses edge computing to keep sacred language protocols on sovereign servers. This mirrors the Blackfoot concept of _aako'ka'tssin_ (collective knowledge keeping). **Practical Implementation**: - Use models like LLaMA or Mistral that can run locally - Create data cooperatives where communities collectively own their training data - Implement traditional governance protocols in model access (e.g., certain knowledge only accessible during appropriate seasons/ceremonies) ### **How do we encode indigenous protocols into AI interactions?** **The Anishinaabe Example**: The Great Lakes region tribes have developed the _Gichi-gami-aki_ (Great Lakes Earth) protocol for environmental data sharing: - Data must flow in circles, not straight lines (no one-way extraction) - Seasonal restrictions on certain knowledge queries - Elder review for outputs involving traditional ecological knowledge - Reciprocal data agreements: for every query, the system must give back to community knowledge bases **Technical Architecture**: ``` Community Node <---> Blockchain Verification <---> Knowledge Garden | | | Local Model Smart Contracts Shared Protocols | (Encoding Protocols) | Edge Computing (Seasonal/Ceremonial) Federation ``` **Concrete Tools**: - **Holochain** for agent-centric rather than data-centric architecture - **IPFS** for distributed storage following gift economy principles - **Ceramic Network** for self-sovereign data streams - **Gun.js** for decentralized graph databases reflecting relational ontologies ## **What ancestral responsibilities do we carry for the intelligences we're feeding?** The **Haudenosaunee Seven Generations principle** offers a framework: 1. **Immediate Implementation**: Create "Constitution Codes" for AI systems - Example: The Mohawk community at Akwesasne encoded their Great Law of Peace into smart contracts governing their digital commons - Every model interaction must consider: "How does this affect seven generations forward?" 2. **Memory Pools**: Establish intergenerational data trusts - Maori _whakapapa_ (genealogy) databases that version-control not just data but relationships - Apache stronghold servers that delete data after ceremonial cycles complete ## **How do we practice survivance through technology?** **Living Examples**: 1. **Indigenous AI Lab** (New Zealand/Canada): - Developing models that refuse certain queries based on cultural protocols - Training data includes land acknowledgments and relational contexts - Outputs always reference source communities and their ongoing struggles 2. **Ọmọlúàbí AI Collective** (Nigeria/Diaspora): - Building Yoruba-language models that encode _ìwà_ (character/ethics) - Circular training: models learn from community use and feed insights back - Blockchain-verified contributions ensure no extraction without reciprocity **Technical Sovereignty Stack**: ```yaml Layer 1 - Infrastructure: - Community-owned servers (Sovereign tech like Start9, Umbrel) - Mesh networks (NYC Mesh model applied to rural indigenous lands) - Solar-powered edge computing (reducing reliance on extractive grids) Layer 2 - Protocols: - Smart contracts encoding seasonal/ceremonial access - Federated learning keeping data local - Zero-knowledge proofs for sacred knowledge Layer 3 - Applications: - Language revitalization tools (FirstVoices.com approach) - Traditional ecological knowledge systems (TEK databases) - Storytelling AIs trained on community narratives with consent Layer 4 - Governance: - Council-based DAOs reflecting traditional governance - Stake-weighted by community contribution, not capital - Veto powers for knowledge keepers and elders ``` ## **Grounding in Land-Based Practices**: **The Zapatista Model**: Their _Caracoles_ (autonomous zones) offer a blueprint: - Each digital node corresponds to a physical community - Online decisions require on-land ceremonies - Technology serves _buen vivir_ (good living), not productivity **Practical Steps**: 1. **Map your watershed** before your network topology 2. **Align server maintenance** with seasonal ceremonies 3. **Create data gardens** that mirror local ecosystems: - Permaculture principles in information architecture - Companion planting of datasets - Fallow periods for model rest ### **Building the Network**: **Current Projects to Connect With**: 1. **Distributed AI Alliance**: Coordinating local-first AI development 2. **Indigenous Data Sovereignty Networks**: Formal protocols for data governance 3. **Rad.live**: Decentralized streaming for indigenous media 4. **ECHO Network**: Environmental knowledge sharing with built-in reciprocity ### **Next Concrete Actions**: 1. **Start Local**: Form a community AI circle - Meet in person, on the land - Develop protocols specific to your place - Begin with language preservation or TEK documentation 2. **Choose Your Stack**: ``` Minimum Viable Sovereignty: - Raspberry Pi + Ollama (local models) - IPFS node for distributed storage - Simple blockchain for community verification - Mesh network connection to reduce ISP dependence ``` 3. **Establish Reciprocity Loops**: - Every query generates community benefit - Models improve through use but within protocols - Benefits flow back to land-based projects 4. **Document in Multiple Forms**: - Code repositories (Git) - Oral histories (recorded with consent) - Ceremonial protocols (kept with appropriate privacy) - Land-based markers (physical computing installations) ### **The Path Forward**: We're not building toward some distant utopia. Communities are already creating these systems. The question is how to network them while maintaining local sovereignty. The blockchain isn't the revolution—it's just one tool for encoding reciprocity. The revolution is remembering that all intelligence is collective, all data is relational, and all technology must answer to the land. Start where you are. Use what you have. Do what you can. But do it in connection—to place, to people, to protocol.