In the MyData Matters blog series, MyData members introduce innovative solutions that align with MyData principles, emphasising ethical data practices, user and business empowerment, and privacy.
High-quality journalism is critical to democratic society and, increasingly, the foundation of reliable AI systems. Yet newsrooms worldwide face an existential paradox: the very content that makes AI systems more capable, trustworthy, and safe is being harvested without consent, attribution, or compensation – undermining the financial viability of the organizations that produce it. We now risk a genuine collapse in the information ecosystem if we don’t create new models to support it.
In recent years, AI companies have approached individual media organizations with seemingly-generous licensing deals. But those AI companies bargain from a position of strength: they can get similar data elsewhere, or ingest it unilaterally with manageable legal risk. On the other hand, media organizations who accept these deals do so from a position of weakness. If they walk away, they might get nothing, forgoing a sustainable source of revenue.
Many of the challenges facing information producers can be understood from the fact that they are siloed into many separate platforms, while the AI products that compete with them can integrate all of their information into a single place. Without similar efforts at upstream integration, they risk being “divided and conquered,” which could result in extreme economic challenges.
One effort at such integration is being led by OpenMined Foundation and RadicalxChange Foundation (2024-2025 MyData Award winner), which have developed a new prototype empowering creators to manage AI training on their content collectively, instead of individually offering their data on a fragmented basis.
The way it works is simple: news outlets install open source software connecting their content to a secure endpoint managed by a representative, such as a trade association. This enables them to provide a value-added, combined product offering and thus work for their shared interests against a few extremely well-financed AI companies. Think of it as trade unionism for the digital age – creators retain control while gaining collective bargaining power.
This is very aligned with the European Commission’s recent embrace of data unions, which we hope becomes Europe’s comparative advantage in AI. Its data spaces, such as the Trusted European Media Space (TEMS), already provide infrastructure for this kind of sectoral coordination. Similar initiatives in the Global South, like CTRL+J in Indonesia, demonstrate the model’s potential across different contexts.
The MyData Global community has long championed updating approaches to data governance, and it is becoming clear that aggregated data access points and attribution layers are an important part of the AI economy. However, most actors providing these services are private startups that extract a large cut of revenue and are likely to get acquired by larger companies or investor consortia, meaning they are not credibly aligned with the interests of information ecosystems.
To our knowledge, this is the only project building free software that allows information producers to maintain control while monetizing their content through an aligned intermediary.
The tools are open source. The precedents are emerging. The question is whether we’ll use them before it’s too late.
ABOUT RADICALXCHANGE FOUNDATION
RadicalxChange Foundation is a 501(c)(3) nonprofit organization dedicated to democratic innovation and institutional design through operational partnerships and experimental projects between academia, government, art, technology, and beyond.
Authors:
Malik Lakoubay, Director of Policy & Outreach, RadicalxChange Foundation (LinkedIn)
Jack Henderson, COO, RadicalxChange Foundation (LinkedIn)
Matt Prewitt, President, RadicalxChange Foundation (LinkedIn)