✘ Creatively imagining the future: Reporting from the Mumbai Convention, 2034
For the protection of creators, publishers, authors, and distributors
Hi everyone,
This week, we (Maarten & Bas) teamed up for some time-travelling, inspired by the book How to Fall in Love with the Future by environmental activist Rob Hopkins.
Our time machine landed us in 2034, allowing us to shape our opinion and vision on solving some of the problems we’re facing in 2026. Our future vantage point provides a unique glimpse of what the rise of artificial intelligence since the mid-2020s has meant for the music industry and for those who do not feel properly represented by it.
Love,
Bas & Maarten
The 2034 Mumbai Convention for the Protection of Creators, Publishers, Authors, and Distributors is the most important cultural policy moment since the original Berne Convention of 1886. Some see it as an overdue correction to a system that has been broken and failing musicians for decades.
The delegates filing into the Jio World Convention Centre look like people who have been arguing for a very long time. And they have. The discussions here have their roots in a crisis that has been building since at least 1999. To understand how we got here, we’ll focus on three strands that weave their way through the Accord about to be signed.
The long acceleration
The processes that made artists anxious had been running for years before generative music startups Suno, Udio, and others arrived in the 2020s: the attention economy’s relentless incentive to produce more content faster, coupled with a democratization of production and distribution of creative work.
The traditional idea of authorship had been scrambled by sampling and remix culture. Meme culture had taught audiences that cultural objects belong to whoever uses them. Algorithmic recommendation had produced a hyperspecific, fragmented audience landscape where mass media dominance was giving way to thousands of micro-cultures simultaneously. Generative AI poured fuel on all of these dynamics at once.
The major labels watched this happen and responded. The Spotify era taught them that you could turn a piracy crisis into a licensing revenue stream, provided you owned enough catalog to sit at the table. The same logic applied to AI-driven music generation platforms: sue to establish leverage, then license on terms you control.
By 2026, all lawsuits had settled and licensing deals were in place. These deals made short-term business sense, like the rise of investment funds treating song catalogs as asset classes, but solved nothing for working musicians whose recordings had trained the models.
As the gap between institutional compromise and artist reality grew in the late 2020s and early 30s, the scene for the Mumbai Convention was set.
The penny fraction problem
The dominant proposal for compensating rightsholders whose work fed training data was basically the pro rata streaming model, but aimed at attributing which recordings influenced a given AI output.
Since generative models do not work like sampling machines, attribution-based models showed us that if you cannot meaningfully calculate what fraction of an AI model's value came from any particular artist's contribution to the training data, the answer is that the model's value derived collectively from the entire corpus.
That raised a different question: what would it mean for the corpus contributors to own a share of the platform itself, rather than a share of each individual output?
“We stopped asking artists to be entrepreneurs. We made the system carry the weight.”
Under what delegates here are calling the Corpus Equity Principle, platforms that derive commercial value from AI models trained on creative work must offer contributing artists equity stakes in the platform. The valuation questions have proven hard and the geopolitical complexity of enforcing equity obligations across jurisdictions with such wildly different corporate law frameworks consumed many hours of negotiation. But the principle itself stands.
Creative work has earned the right to be treated as a form of capital investment that deserves a capital return.
What the market could not fix
What has changed since 2028 are the political and economic conditions for intervention. The scale and speed of AI’s impact on creative labour created a coalition that had not previously been able to cohere into a meaningful form.
Independent artists and label groups who had survived the streaming age
People who came of age through streaming and generative AI
Musicians in the Global South whose work had been systematically undervalued in countries with strong traditions of collective rights management
The Mumbai Convention is the product of this coalition, backed by a broad assembly of transnational organizations and, critically, governments in South Asia, West Africa, and Latin America. This coalition saw how AI models hungry for training data were a new version of a very old story about whose cultural resources get extracted.
“The Corpus Equity Principle means we've moved into monetization models that benefit all artists.”
The Berne Convention of 1886 was a European project, designed to protect European authors from European publishers in an era when Europe defined what counted as culture worth protecting. Despite a strong sense of power imbalances and compromise, the 2034 Mumbai Convention assembled a genuinely global forum with musicians at the table for the first time in 150 years of international copyright history.
This newsletter was supported by Imagine Intel, the zine and cultural artifact published by Mozilla Foundation to accompany the Imaginative Intelligences Assemblies work led by the Foundation’s Creative Futures program. Check out the work of The Counterstructural Commons, Mozilla Foundation’s cultural R&D residency at Rhizome.
LFG! 🫡