community data stewardship and alternatives to Big AI
From 2022, I developed a number of projects on community data stewardship as alternatives to mainstream AI alongside researcher Jennifer Ding.
As the community manager of The Turing Way, it became increasingly clear that grassroots data practitioners were feeding into much bigger systems, least of all huge projects like OpenAI’s Chat GPT.
In discussions with Jennifer, we started drafting what felt like a . She was collaborating with people building alternative LLMs at places like HuggingFace, CohereAI and others.
You can read more about the outcomes of this work in a few talks, workshops, and publications:
- Towards Openness Beyond Open Access: User Journeys through 3 Open AI Collaboratives” (2023)
- “Community Governance Opportunities in the ML Pipeline” (2024)
- “Participatory Data Stewardship in AI” (2024)
As a part of this work we also presented this work in a few different locations, including: “Designing for Data Rights in the AI Production Pipeline” session at the Mozilla Festival in 2023 and “Participatory data stewardship in AI” at Mozilla’s Data Futures Lab’s Community Call.
The Miro board we developed and have used as refence is linked here.
There are also a chapters about this work with The Turing Way handbook: Data Governance for the Machine Learning Pipeline.