The idea that personal data is the new oil has become so familiar that almost no one questions it anymore. Companies extract it, refine it, and sell it. Google, Meta, and Amazon then appear as prospectors who discovered a resource with no prior value. They built the necessary infrastructure and generated wealth where none existed before. The exchange—free services in return for data—seems reasonable enough.
This story contains some truth. Platforms did create real structures. Their models solve information problems that no single user could tackle alone. When a search about medical symptoms returns useful results, that happens because millions of previous searches fed the trend. There's a collective component to that value, even if the final profit is private.
Still, this version leaves out a structural blind spot. You only need to look at what happened to common lands to recognize it.
Between the 16th and 18th centuries, the English Parliament passed the Enclosure Acts. They promised improved productivity through private management. In practice, they fenced off and privatized millions of acres that rural communities had managed collectively for generations. Peasants lost customary rights to grazing, gathering firewood, and fishing. They didn't lose these because of inefficiency. They lost them because the law declared those rights nonexistent and favored landowners instead.
The parallel with data isn't poetic. It's structural. We generate personal data through our behavior, our relationships, and our fears. It constitutes a collective good in a precise technical sense. Legal frameworks that grant ownership to whoever captures the data replicate the same enclosure logic: they turn the commons into private property through legal fiction. Peasants at least could see the fences. We don't even get that.
What's interesting is that those commons weren't chaotic free-for-alls. The ones that survived had clear local rules about who had access, how much they could take, how decisions were made, and what happened when rules were broken. Elinor Ostrom documented these cases for decades and showed that well-governed commons don't inevitably fall into the tragedy Hardin predicted. They collapse when the institutions sustaining them are destroyed. Which is exactly what the Enclosure Acts accomplished.
Data cooperatives are trying to rebuild those institutions for the 21st century. Not as metaphor, but as concrete architecture. Members collectively agree on what data they contribute, under what conditions it gets shared, and how benefits get distributed. The cooperative acts as a trusted intermediary. No outside actor gains access without the group's consent.
Real implementations exist. MIDATA in Switzerland lets citizens pool health data and collectively decide which researchers can use it and under what rules. Driver's Seat Cooperative in the United States collects data from gig-platform drivers and negotiates it with cities, returning value to the people who generated it. These cases show the architecture is viable.
That architecture faces several problems at once. Verifying membership and voting rights without creating vulnerable central points. Controlling access with granularity. Distributing benefits transparently and auditably. Decentralized identity designs and zero-knowledge proofs offer partial paths forward. Their actual integration into cooperative structures remains experimental. No consolidated standards exist. I'm not sure how to resolve the tension between growth capacity and local control without introducing new fragilities.
The debate over artificial intelligence systematically avoids the question that matters most. It focuses on technical capabilities and timelines for supposed AGI. It ignores who owns the data that trains these models. Large models are, literally, distillations of collective human intellectual labor: texts written by people who never consented to this use, images created by photographers who received no compensation. If data is the fuel, digital enclosure isn't history. It's happening now, on a planetary scale.
Analysts who follow this space warn that without solid legal frameworks, cooperatives risk being absorbed or copied by dominant platforms. A large company could launch its own façade version, adopting the language of collective empowerment without ceding any real control. The risk of image-laundering is structural. It's already happened with other community initiatives.
Even so, institutional memory of how commons functioned survived the enclosures. It persisted in places that kept their local rules alive. That's precisely what platform-cooperativism frameworks are now trying to codify: in protocols and bylaws that make it technically difficult, not just illegal, for a centralized actor to capture the value.
Interestingly, the most solid data cooperatives didn't start as tech projects. They emerged from communities facing concrete governance problems. Patients frustrated at not controlling their own medical records. Drivers who wanted to understand how algorithms determined their earnings. Technology arrived afterward, in service of a need that had already been articulated.
The Enclosure Acts stretched across two hundred years. Digital enclosure has taken barely a few decades.
Will we choose this time to build the institutions that protect the commons before they disappear entirely?
