Abstract
Communities, researchers and organisations worldwide are being presented with a false choice: surrender their data to consentless scraping, sign exclusive deals with megacorporations, or abandon the digital future altogether. This workshop introduces a radical alternative: a participatory data ecosystem where communities can actively shape how their data powers AI. Rather than treating data as a resource to be extracted, the Collective treats it as a community asset, grounded in authentic consent, retained ownership, and collective governance. We explore how these principles can realign the incentives behind AI development — shifting data practices from exploitative extraction to genuine empowerment. The 45-minute session is designed as a highly interactive dialogue. Participants will first engage in a short, staged conversation with facilitator from the Mozilla Data Collective about the harms of current AI data pipelines. Then, through small group activities, they will map out their own communities’ data flows, identify risks, and experiment with alternative governance models. We will collectively surface shared challenges and co-develop principles for ethical, community-led data infrastructures. The workshop will close with an open debate on scaling these models: what legal, technical, and social mechanisms are needed to embed consent and ownership into AI’s data foundations? An article summarizing outcomes will be developed. This session offers a hands-on opportunity to rethink how human-centred AI can emerge from human-centred data.
Presenters
E.M. Lewis JongFounder & VP - Mozilla Data Collective, Mozilla Data Collective, Common Voice, Mozilla Foundation, California, United States
Details
Presentation Type
Theme
2026 Special Focus—Human-Centered AI Transformations
KEYWORDS
Collective Data Governance, Community-led AI, Multilingual AI
