2026-07-09 –, AI Barn
Mainstream AI models often ignore or exploit marginalized and stateless languages. In this interactive workshop, we will explore the grassroots blueprint for building sovereign, low-resource language AI models—from ethical data collection to offline deployment in rural communities.
The current artificial intelligence boom is leaving marginalized, stateless, and deep-rural communities behind. Massive, centralized LLMs are trained heavily on dominant languages, while low-resource languages face digital erasure. When Big Tech does attempt to include these languages, the data extraction is often exploitative and disconnected from the community's actual needs. To uphold the DWeb principles of Humanity and Distributed Benefits, we must empower local communities to build, own, and govern their own localized AI.
What We Will Do
Drawing on my experience founding eimiAI, this 60-minute interactive workshop will guide participants through the realities of building AI from the bottom up. We will dedicate the majority of our time to hands-on problem solving:
Part 1: The Grassroots Blueprint (15 mins): I will share how eimiAI approaches the technical and social challenges of low-resource AI. We will cover community-led data generation, privacy-first data structuring, and selecting lightweight models that can run on constrained hardware.
Part 2: The Sovereign Data Challenge (35 mins): Participants will break into small groups. Each group will be assigned a hypothetical stateless community with a unique language and severe infrastructure constraints (e.g., no internet access, low digital literacy). Groups will be challenged to design an ethical data-gathering campaign and map out a local-first tech stack to deploy a basic translation or voice-to-text model for that community.
Part 3: Group Debrief (10 mins): We will reconvene to share our frameworks, discussing how DWeb protocols can better support sovereign, offline-first AI.
Why This Matters
This session moves beyond theoretical AI ethics into practical survival. It centers the voices of those not at the table and equips DWeb builders with the methodologies needed to ensure AI acts as a tool for language preservation and human agency, rather than just another centralized monopoly.
Michael Suantak is the founder of ASORCOM (Alternative Solutions for Rural Communities) and eimiAI. Operating at the intersection of indigenous rights and decentralized technology, he engineers autonomous, offline-first communication networks in extreme conflict zones, specifically along the India-Myanmar border.
During state-sponsored internet blackouts and kinetic warfare, Michael’s work focuses on keeping community knowledge, medical coordination, and human agency alive. He is the architect of the "Digital Tarzan" philosophy, which advocates for frugal, rugged infrastructure built from the soil up. His field deployments combine 12V DC-native solar power, tactical LoRa mesh networks, HF radio, and localized edge AI to ensure marginalized communities retain cognitive and digital sovereignty when the centralized grid fails.
Michael believes that true autonomy is not given—it is engineered. His ongoing mission is to help rural populations transition from vulnerable digital consumers into resilient digital sovereigns.
- Resistance, Resilience and Regeneration - The Role of Collective Infrastructure in Times of Collapse
- The Dark Horizon Simulation: Architecting Autonomous Comms in an Extreme Conflict Zone
- Cognitive Sovereignty at the Edge: Co-Designing Offline AI for Conflict Zones
- Beyond the Canopy: Building Local Networks & Marginalized Public AI at the Rural Edge
- The Ancient Signals Treasure Hunt: Pre-Physics Communication Models & Optical Cryptography