Experiments¶
This research employs a series of iterative experiments to explore collaborative AI frameworks. Each experiment builds on previous learnings while testing different approaches to community-governed artificial intelligence.
The experiments follow a Research through Design methodology, where each prototype serves as both a technical proof-of-concept and a probe to understand community needs and governance challenges.
Key Research Questions Explored:¶
- How can communities maintain meaningful control over AI systems?
- What interfaces make AI governance accessible to non-technical users?
- How do different deployment strategies affect community autonomy?
- What are the trade-offs between simplicity and functionality?
LAIA Project
Local AI for All - 2024
Neighborhood intelligence platform using AI and crowdsourced data. Demonstrated potential of local AI solutions while highlighting challenges of language barriers and data scarcity. Collaboration with Nuria Vallès.
Slack Workspace
Platform Integration - 2024
Testing AI integration in existing community platforms. Revealed complexity of dependencies and API management, leading to questions about maintainability and community autonomy.
BLOB (Blob Browser)
Community-Driven Paradigm - 2024
Dramatic shift from city-scale to individual/community scale. Enables users to run AI models locally, collect data, and contribute to decentralized networks. Focus on energy consumption and infrastructure requirements.
LLUM 2025
Local AI Installation - 2025
Interactive exhibit collecting and showcasing local, public data. Provided insights into data collection and presentation while raising concerns about energy consumption and scalability of running GPT and DALL-E on public installations.
Oatflake
Federated Knowledge Systems - 2025
Lightweight approach to local AI deployment, focusing on reducing technical barriers for community adoption. Enables cross-community exchange and builds upon learnings from previous experiments.
Research Roadmap
Scaling Strategies - 2025-2030
Future directions for community-governed AI ecosystems. Long-term vision for facilitating cross-community exchange, enabling users to build upon and learn from one another's local intelligence.
Experimental Journey¶
Inspired by projects like OpenStreetMap, Aïna, and Oio News, this exploration of local AI implementations follows an iterative path from neighborhood-scale solutions to federated community networks. Each experiment builds on previous learnings while testing different approaches to community-governed artificial intelligence.
The journey demonstrates a clear evolution: from city-scale approaches (LAIA) to public installations (LLUM 2025) to individual/community-focused frameworks (BLOB) and finally to federated deployment systems (Oatflake). This progression reveals the importance of iterating, adapting, and rethinking AI implementations to better suit local needs, constraints, and aspirations.
Each experiment contributes to a growing understanding of how communities can meaningfully govern their own AI technologies while maintaining autonomy and reflecting local values.