The current AI innovation paradigm is driven by large-scale extraction and (mis)appropriation of creative and cultural resources into proprietary AI systems. As AI systems increasingly mediate cultural production and information access, questions of who controls these knowledge infrastructures have become increasingly urgent. These developments raise serious concerns not only for the livelihoods of artists and creators but also for the collective cultural and developmental rights of communities and states, particularly in the Global South.
This issue brief, ‘Governing AI for the Cultural Commons: Beyond Intellectual Property,’ aims to surface the emerging challenges posed by the dominant AI paradigm by foregrounding the collective dimensions of the public domain and cultural data as critical developmental infrastructures. It examines how current AI innovation pathways are reshaping cultural production and knowledge systems, analyses the role of IP frameworks in structuring these transformations, and explores emerging governance directions towards safeguarding epistemic justice and enabling fairer value sharing in AI-driven economies.
This issue brief was collaboratively developed by the AI, Culture, and Intellectual Property Subgroup of the UNESCO Global Civil Society Organizations (CSO) and Academic Network on AI Ethics and Policy. IT for Change served as the co-lead of the subgroup alongside AI Impact Alliance and anchored the issue brief with contributions from subgroup members. The brief was informed by a combination of desk research and stakeholder consultations, including an expert roundtable and regional focus group discussions across Africa, Asia-Pacific and the Ibero-American region.
Key insights from the brief
- Dominant AI systems rely on large-scale extraction from cultural commons and public domain resources, placing pressure on creative labour, public sharing practices, and the conditions that sustain the commons itself.
- AI systems increasingly shape how cultures are represented and reproduced, often flattening cultural meaning and reinforcing homogenised representations detached from social and historical context.
- While cultural resources and labour from across the world feed AI systems, control over AI infrastructures remains concentrated in a few hands, reinforcing technological dependence and limiting the ability of Global South countries and communities to shape AI systems in line with their own priorities.
- Current IP frameworks play a central role in enabling these dynamics. Copyright exceptions facilitate large-scale data extraction, while trade secrets, patents, and contractual arrangements are increasingly used to consolidate proprietary control over AI systems and their outputs.
- Conventional IP frameworks remain poorly suited to the AI paradigm, particularly where knowledge production operates through aggregation and abstraction at scale and where cultural production is collective, incremental, and socially embedded.
- Conventional IP frameworks are ill-equipped to address the new paradigm of knowledge and cultural production shaped by AI, particularly where systems operate through aggregation and abstraction at scale.
- While IP reforms remain important, broader governance approaches are needed to reclaim the knowledge infrastructures of tomorrow. These include stewardship-based data governance, public AI infrastructures, global frameworks for AI training data governance, and redistributive mechanisms to support cultural and knowledge commons.
Read the issue brief here.