The first question we brought home from the AI Bootcamp, hosted by LegaCoop and ICA in Rome was how the cooperative economy can find its place across the whole AI value chain. The next question is where P6 fits—and how our work could sit alongside the many other things cooperatives might build.

We see P6 as one contributor to that wider ecosystem. We bring the architecture that lets cooperatives capture and store their own knowledge, keep control of it, and share what they choose with the wider cooperative economy.

There is room to build throughout the system

AI depends on minerals and hardware, energy, data centres and networks, as well as data, models and the applications people use. Cooperative enterprises could have a role at many of those points.

Cooperative hosting, shared computing resources, models and tools for particular sectors could each contribute. Different initiatives would bring different experience and solve different parts of the problem.

We may not have the resources to immediately build a cooperative model that competes with the largest technology companies. But we can still explore cooperative ownership and democratic control in the parts we are able to build. Work in one part could make the next step possible elsewhere.

P6’s contribution is how knowledge and tools connect

When we talk about architecture at P6, we mean the software and rules that help cooperatives build and use their own body of knowledge. Who can add to it? Which tools can consult it? What can be shared, with whom and for what purpose?

P6’s approach is model-agnostic: the architecture is designed to work with different AI models. A cooperative’s knowledge and sharing rules should not depend on committing to one model forever. That is a design approach, rather than a claim that every model is already supported.

There is a useful distinction here. A model learns patterns during training. An application can also give it relevant knowledge when someone asks a question. A cooperative could update a member guide, for example, and make the new version available to its tools without retraining the underlying model.

P6’s focus is ownership and democratic control of that knowledge component. Cooperatives could build collections of their own documents and experience, keep sensitive material private, and share selected resources across federations, sectors and the wider economy. Allowing a tool to consult a document is a different permission from allowing it to be reused for model training.

Different initiatives can make each other more useful

Imagine a housing cooperative using a sector-specific tool run by a cooperative technology provider. It could manage its own knowledge and consult resources shared by its federation, using a model to provide the underlying AI capabilities.

A P6 approach to knowledge and permissions could be one contribution, alongside the hosting, the application and sector expertise.

A useful ecosystem would leave room for different solutions and let cooperatives choose tools that suit them. Connections could make those independent contributions more useful.

The connections need cooperative rules too

The organizations involved would need shared rules for moving information between services, managing permissions and handling mistakes.

Contributors need a say in changes and a fair return for what they share. Cooperatives should be able to move their knowledge or change providers.

Outside models or infrastructure may still be involved. Gaining control over knowledge is meaningful, while leaving other dependencies to understand and address.

Build from the contributions we can make

P6 wants to develop and test its contribution alongside others. As different cooperative initiatives learn to work together, they could become more useful than they would be alone.

Join the discussion

What could your cooperative contribute to a wider AI ecosystem, and what would help it work with others? It might be knowledge, a service, infrastructure or experience from your sector.

Explore P6 and CoopAI

New to P6? Read the CoopAI overview or watch the CoopAI Explainer to explore our approach.

Read our Presentation Deck.