Generative AI is reshaping how knowledge is produced, used, and shared—and Wikimedia sits at the center of that shift. This paper aims to understand the ongoing structural transformation of the knowledge ecosystem and to help articulate a direction for the Wikimedia Movement’s response.
The new AI knowledge loop
This white paper traces the history of the online knowledge ecosystem through three successive knowledge loops: the first tied to search engines, the second to social media, and the third connected to emerging AI technologies.
The third loop breaks the reciprocal dynamics that sustained Wikimedia. AI systems now consume the commons at a massive scale without returning the value that keeps them alive. The result is not only a sustainability challenge driven by infrastructural strain but also a threat to Wikimedia’s ability to remain an active force in the production and governance of knowledge. This paper describes the risk of Wikimedia becoming a “dark commons”: widely used yet structurally invisible and lacking agency—an unseen substrate feeding proprietary systems.
Solving the Paradox of Open for Wikimedia
To remedy this issue, the paper issues a call to action: transition from a defensive posture toward generative AI to one focused on generating value by contributing to the discourse and governance around these systems. It examines the Paradox of Open, which is how open sharing can enable extraction rather than equity under conditions of extreme power asymmetry. Building on previous discussions with community members, movement leaders, and outside experts included in our “Collective Intelligence vs. Artificial Intelligence” report from the Wikimedia CH Roundtable on Wikimedia and AI, the new white paper maps three types of value flow (extractive, reciprocal, and generative) and three strategic postures available to stewards of the commons: defensive closure, passive acceptance, and an active stance. The two most common responses today—closing access to protect the commons or maintaining openness in the expectation that reuse benefits everyone—leave structural problems unaddressed.
Wikimedia AI Mission
From this diagnosis, the paper proposes a Wikimedia and AI mission for 2030: to work in coalition with public AI developers and commons-aligned partners to build and govern the interfaces and infrastructures through which the knowledge commons are produced and reach readers. The aim is for the Wikimedia Movement to remain one of the primary user-facing knowledge interfaces, with agency over how knowledge is produced and used in the AI era.
The mission has two parts:
- Pro-commons governance—reshaping norms around attribution, reciprocity, and the protocols that structure the knowledge ecosystem.
- A Wiki AI technical stack—supporting the development of commons-aligned data and application layers, and contributing to public AI model development.
Supporting these are a mission network that coordinates action across movement entities, volunteer editors, and aligned partners, and dedicated mission investment to resource the work.
The paper, written by Alek Tarkowski from Open Future and Ilario Valdelli from Wikimedia CH, does not represent any single organization’s position. It offers an emerging direction and a shared language for various people and entities that want to support this mission. It is intended as a contribution to a broader conversation within the Wikimedia movement and in the broader digital commons space: a starting point, not a conclusion, and an invitation to build a coalition ready to pursue a shared Wikimedia and AI mission.
