A governance architecture for high-complexity societies — a designed social technology that captures individual data across three dimensions (physical, virtual, cognitive), converts it into structured labels via tensor factorization, and coordinates collective action through stigmergic traces and smart contracts — without concentrating power or reducing human diversity.
The Problem
Contemporary society faces a structural condition that existing institutions were never designed to handle: complexity. Interdependence, diversity, speed, scale, and non-linearity define our world — but our dominant governance model, centralized representation, was designed for simpler contexts.
Representation becomes a bottleneck: too many interdependencies, too many perspectives, too much information to process through a small number of centralized decision-makers. This is not a failure of specific politicians or parties. It is a structural mismatch between the complexity of the society we have and the governance technology we inherited.
The Core Idea
Governance — institutions, rules, representation systems, decision procedures — is not a fixed natural category. It is a social technology: a human-designed artifact for coordination. Like any technology, it can be studied, prototyped, and redesigned when the context changes.
This reframing is sOS's foundational move: from “who governs?” to “how do we govern?” — treating governance as a scientific and engineering problem, not only a political one.
The Data Pipeline
1. Data Capture (3 Dimensions). Each individual generates data across three dimensions — Physical (GPS, mobility), Virtual (digital activity, preferences), and Cognitive (physiological state, emotional responses via EDA sensors). Solid Pods collect this data under the individual's sovereign control.
2. UILL Storage. The Unique Information Library Ledger stores each person's multi-dimensional data with cryptographic provenance and consent controls. No platform, government, or corporation owns this data.
3. Tensor Factorization. A mathematical model transforms the 3-dimensional data into a structured label space. This is where raw data becomes computable governance information.
4. JSON-LD Labels. The output is structured labels using the sOS vocabulary — encoding capacities, needs, preferences, history, and relations between individuals.
5. Stigmergic Traces. Labels are published as trace data into the Social Machine — a shared information environment where agents (human and AI) read, interpret, and respond to patterns, just as ants read pheromone trails.
6. Smart Contracts (The Hybrid Bridge). Six contract types — Propose, Validate (Labels.sol, computing S² variance to detect label interdependencies), Delegate, Execute, Save/Compute, and Finance — form the operational bridge between the human domain (deliberation, judgment) and the machine domain (computation, matching, execution).
Architectural Principles
Multiple agents — individuals, organizations, AI systems — coordinate through shared information environments, without a single controller. Rules can emerge from and adapt to data rather than being imposed once and frozen.
As ants coordinate through pheromone traces in the environment, humans can coordinate through data, labels, and registries. Individual actions leave computable traces that form a shared information environment — enabling coordination without top-down control.
Labels are not simple tags. They are JSON-LD structured data generated through a specific pipeline: individual data is captured across three dimensions (physical/GPS, virtual/digital, cognitive/physiological) via Solid Pods, stored in each person's UILL, and transformed by tensor factorization into computable labels encoding capacities, needs, preferences, history, and relations. Labels are the vocabulary; stigmergy is the process. Together they make social complexity computable for distributed systems.
Each person has a personal AI agent that knows their context, filters relevant information, and helps them participate in governance decisions without requiring infinite attention. The AI serves the person — not the other way around.
Every individual controls their own data through a Unique Information Library Ledger (UILL). Data belongs to people — not platforms, governments, or corporations. Individual data sovereignty is the foundation of collective coordination.
What sOS Is Not
Not a political party. We do not run for office. We collaborate with public administration — we do not compete with it.
Not an ideology. We do not prescribe specific policies. We provide an architecture within which diverse policies can coexist, compete, and evolve.
Not techno-solutionism. Technology does not replace politics. It changes the conditions under which politics operates.
Not a cryptocurrency project. sOS uses token mechanisms strictly as internal coordination and accounting instruments for distributed work — not as speculative assets or a currency. Governance tokens are non-transferable and serve as participation records, not financial instruments.
Not a finished product — but already operational. 11 Smart Contracts v6 are deployed and verified on Polygon Amoy testnet, 16 Working Papers are completed with dual-model review, and autonomous services (Sentinel, Fixer, Health Daemon) monitor and maintain the system 24/7. The architecture is implemented; validation through the Barcelona Sandbox comes next.
sOS Research Lab
A non-profit scientific association hosted at the Universitat de Barcelona. The operational vehicle for the sOS thesis: an organization designed as an experiment in adapting human organization to AI-mediated governance.
sOS Research Lab
Universitat de Barcelona
Avinguda Diagonal 696, office 320
Barcelona
hi@socialbigdata.io