Product research.

A people-controlled Web 3.0 data relationship platform — a bridge between individuals and brands where privacy, choice and value take centre stage. People create, own and share their own accurate personal-data profiles (their zero-party data, each person’s “Profila”), and license — never sell — parts of them through data subscriptions, while controlling the attention brands can pay for.

It rests on three layers: a proof-of-identity layer (a Profila digital ID with Decentralised Identifiers federated to the EU eID, an identity/decision/payments blockchain, homomorphic encryption and a personal data space); data-subscription smart contracts through which people license buckets of data; and multi-party computation plus a Zero-Knowledge Advertising protocol — “The Cookie Killer” — that matches a person to an offer and delivers a private ad without ever exposing them, replacing cookies across apps, browsers and websites.

The six-partner consortium — De Roos Advocaten (coordinator, legal), Universidad Carlos III de Madrid (zero-knowledge advertising), Partisia (multi-party computation), Profila, Publicis Bulgaria and Parhelium — submitted it to Horizon Europe in 2023. It was not funded, but it is the blueprint tying the rest of this work together: the 42-True model, the UC3M zero-knowledge advertising paper, the CERN privacy ontology, the AI Privacy Lawyer, and the Cardano × Partisia MPC project.

The collaboration’s foundational deliverable was a privacy ontology — a taxonomy of the classes and relationships that describe privacy: users and data subjects, entities, policies, operations, transactions, access control and devices. It gave the rest of the work a shared, machine-readable vocabulary for reasoning about personal data.

On that base the team built models and datasets trained on CERN’s own FAQs, operational circulars and rules, a conversational Q&A prototype, and an integrated MVP that let CERN staff ask plain questions about privacy policies and procedures — with the framework published under open-source licenses.

That privacy ontology proved foundational to Profila’s later research and is cited in the 42-True paper. The Q&A prototype is a direct ancestor of the AI Privacy Lawyer.