Safeguarding Information Integrity requires a sustainable AI content licensing model
The battle for information integrity hinges on a fundamental, unresolved question: how do we protect and compensate for the journalism that feeds the global digital ecosystem? As generative Artificial Intelligence models proliferate, a sustainable solution for fair media compensation remains experimental, troublesome and stubbornly elusive.
The current trend of leaving AI platforms to manage the compensation process disadvantages newsrooms in the long term. This critical warning is detailed in a new policy brief by Dr. Courtney Radsch and Karina Montoya, Same Gatekeepers, New Tollbooths: Mapping the AI Content Licensing Market.
The brief reveals that while a small number of large publishers, entertainment companies, and social platforms have secured direct bilateral deals with dominant AI companies, a startup intermediary ecosystem is emerging to connect AI companies with content and data, particularly from news outlets. “However, Big Tech corporations are encroaching into this fragile AI licensing market with a distinct value proposition.”
Their brief is published against the background of AI platforms rapidly cannibalising the human-created content they require to function. Without fair, structured financial frameworks, the economic foundation of independent media faces an existential collapse, threatening the very fabric of public trust and fact-based information.
The Blueprint: The M20 Johannesburg Declaration
Their policy brief tackles threats and opportunities highlighted in the M20 Johannesburg Declaration and associated M20 policy inputs during 2025. Issuing an urgent roadmap for global governance and accountability, the M20 initiative has called on world leaders to enforce and update intellectual property frameworks, ensuring publishers are equitably compensated not just for initial AI training, but also for ongoing data usage in grounding Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG).
The Johannesburg Declaration specifically champions:
- Global South Equity: Investing in media-led local innovation to prevent technological dependencies and protect linguistic and cultural diversity.
- Transparency and Sourcing: Creating robust mechanisms to track information provenance and mandating independent reviews of AI’s impact on news integrity.
- Unified Bargaining: Urging media outlets to collaborate on pricing criteria and lobby for statutory protections rather than fighting isolated battles.
Same Gatekeepers, New Tollbooths
The newly released policy brief demonstrates how AI companies source, value, and compensate for the news and creative content that their systems rely on. The authors note that although a market for AI content licensing is taking shape, the conditions under which it is forming closely resemble those that made prior platform-journalism relationships structurally damaging to the news ecosystem.
The brief was published by the Centre for Media and Digital Governance (CMDG) in the United States – formerly the Center for Journalism and Liberty at Open Markets Institute.
The authors argue that the emerging licensing market is dangerously replicating the predatory dynamics of the search and social media eras. Publishers are trapped in a structural “double bind”: the same dominant tech giants eroding original website traffic are now constructing and controlling the licensing infrastructure meant to distribute alternative revenue. This creates an unequal, three-tiered market consisting of a few exclusive bilateral deals for media conglomerates, a Big Tech-dominated intermediary layer, and a completely uncompensated majority of local and independent creators.
The Threat of “Sloppification”: If AI platforms remain the sole gatekeepers of compensation, the financial baseline for quality reporting will dry up, turning AI into a parasite that ultimately degrades its own data inputs through a slow starvation of original human content.
The report maps the emerging licensing market – and delivers an innovative, actionable agenda for putting AI and content creators on a sustainable, transparent, symbiotic footing before it’s too late. It drew on more than 35 global interviews and consultations with publishers, content creators, and AI licensing startups (with most companies drawn from the US and a couple in Europe).
“The quality of AI outputs depends on an ongoing supply of quality human content. Destroy the economic foundation of that content, and you degrade the intelligence of AI itself. It’s a house of cards built upon quality content as the base,” said Radsch.
“We found major problems with the current licensing market. None are self-correcting. Either we fix this, or we’ll see yet another crisis in journalism and media. We need to build a symbiotic input-output relationship, or AI risks becoming a parasite that kills the content streams it relies on.”
Innovative solutions to avoid systemic failure
The authors propose frameworks to ensure that fair compensation is treated as a legal obligation and democratic necessity, not an act of charity.
The report pairs its diagnosis with a concrete policy agenda designed to equalise the power dynamic between dominant AI firms and content creators, enhance transparency, and restructure the digital incentives that make quality content economically viable. These include:
- Statutory licensing frameworks: Set enforceable, market-wide payment obligations, ensuring all publishers are compensated when AI systems use their content, replacing one-off deals with reliable standards.
- Collective licensing and sectoral bargaining: Explore extended collective licensing and sectoral bargaining frameworks and allow publishers to negotiate as a group with AI firms, enabling smaller and local outlets to secure fair terms alongside larger players. (Note: Collective bargaining has already proven effective in Canada and Australia.)
3; Transparency requirements: Mandate disclosure of deal terms, pricing models, and data usage to correct information asymmetries and support fair market valuations.
- Attribution & compensation at model level: Require technical systems to include bot identification and detect when AI outputs rely on specific sources and distribute compensation accordingly, akin to rights management systems in music.
- Inclusion of local & independent media: Require eligibility and distribution mechanisms to include smaller, regional, and non-English publishers to ensure diversity of data inputs, prevent further concentration of media power, and limit systemic risks.
The policy brief’s mention of the imbalance of power and the need for diversity in language and regions is explored further in an M20 discussion paper by Barbra Okafor, in which she recommends a shared, licensed trust for African media to tackle exploitative data extraction on the continent. Read Okafor’s paper here: A licensed trust for African media data – Ubuntu as a blueprint for Africa’s AI economy
Read Radsch and Montoya’s policy brief here: Same Gatekeepers, New Tollbooths: Mapping the AI Content Licensing Market
ENDS
