Authored by Paul Keller, the whitepaper builds on Open Future’s ongoing work on Public AI and on AI and creative labour, and proposes measures aimed at ensuring a healthy and equitable digital knowledge commons. The paper views generative AI as a transformative cultural and social technology, raising concerns about increasing concentration of control over knowledge and the weakening of institutions that support human-created information. It identifies two key structural risks:
- The dominance of a few powerful actors in AI-driven information access
- The decline of sustainable models for producing reliable information
To address these challenges, the paper calls for the development of Public AI infrastructure and proposes a redistributive mechanism, which would act as a levy on commercial AI systems trained on publicly available content. This kind of funding model would support not only creators and rightsholders, but also cultural heritage institutions, public service media, and open content platforms.
This paper further argues that ensuring a sustainable information ecosystem in the age of generative AI requires structural interventions that go beyond existing legal frameworks. It calls for a new deal: one that recognises the public value embedded in AI systems and ensures that those who produce, preserve, and share knowledge are fairly supported.
The paper develops four key arguments:
- AI’s impact on information is still underestimated—especially regarding control over access (information policy) and the commodification of the knowledge commons (economic sustainability).
- Public AI models and infrastructures are urgently needed, supported through public funding and governed in the public interest.
- A broad understanding of the information ecosystem is essential, one that recognises the diverse actors contributing to the digital commons.
- Economic sustainability requires redistribution beyond copyright, via a levy on commercial AI systems to channel revenue back to those who sustain the information ecosystem.
