European City Squared: Computational Social Science Simulation for Democracy

Project Information

EuropeanCity2

Grant agreement ID: 101178170

DOI

10.3030/101178170

EC signature date 1 October 2024

Start date 1 January 2025

End date 31 December 2027

Funded under

• Culture, creativity and inclusive society

Total cost

€ 2 972 396,25

EU contribution

€ 2 972 396,25

Coordinated by

AARHUS UNIVERSITET
Denmark

Periodic Reporting for period 1 - EuropeanCity2 (European City Squared: Computational Social Science Simulation for Democracy)

Reporting period: 2025-01-01 to 2025-12-31

Summary of the context and overall objectives of the project

Democracy is an evolving institution. The EuropeanCity2 (EC2) project is a Horizon project that responds to a simple but urgent problem: democracies need collective decisions that are both legitimate and effective, yet many current decision procedures discard information about how strongly people care. When every preference is reduced to a single tick, we lose nuance; when we add nuance, we often add complexity. Classic results in voting theory also show that no voting rule can satisfy all desirable fairness criteria at once (Arrow's impossibility theorem). The practical implication is that design choices create trade-offs - an, and those trade-offs should be tested empirically, not merelyted.

The project builds an evidence-generating "democracy laboratory" in software. We develop an agent-based simulation environment - many interacting virtual citizens in realistic settings - so that alternative voting and participation mechanisms can be explored safely before they are tried with people. A central focus is Quadratic Voting, where casting additional votes on an issue becomes progressively more costly, allowing participants to signal intensity ("I care a lot about this"). We also explore extensions that use modern AI models of decision-making and, experimentally, quantum computing ideas for preference aggregation. The pathway to impact is: create the scientific foundations and modular software, run controlled simulation studies on real and privacy-preserving synthetic data, validate insights through pilot studies with partners in Denmark (including the city of Aarhus) and Switzerland, and translate the results into practical guidance. The social sciences and humanities are integral: they set the normative benchmarks (fairness, legitimacy, consent), interpret outcomes in their institutional contexts, and keep the technical work aligned with democratic values.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During the first reporting period (2025), we built the core scientific and technical infrastructure needed for later large-scale simulations and pilots.

We first established a shared baseline on voting systems. This included a systematic review of major electoral and participatory methods and an explicit set of evaluation criteria: fairness, representation, resilience, accountability, comprehensibility, and consent. This work clarifies where intensity-based mechanisms like Quadratic Voting can improve expressiveness, and where strategic behaviour (e.g. collusion or gaming) remains a risk.
In parallel, we advanced the project's quantum strand by implementing and testing a quantum voting rule ("Quantum Majority Rule") on a simple but real quantum circuit. This provides concrete evidence about algorithmic feasibility under realistic hardware noise, rather than treating quantum voting as a purely theoretical exercise.

On the simulation side, we designed a modular architecture that cleanly separates: the population (who votes), the environment (what is being decided), and the aggregation rule (how votes become a decision). We also produced initial data specifications and reusable processing pipelines, including a small voting web application, a workflow for preparing text-based media data, and an agent-based modelling codebase that already supports comparisons between standard voting rules and quadratic voting. Finally, we demonstrated a proof-of-concept pipeline for generating privacy-preserving synthetic voter profiles using an AI language model, and connected it to an interactive visualisation and replay tool so that simulation outcomes can be inspected step by step.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Even in its first year, EuropeanCity2 has begun to move beyond the state of the art by turning ideas about "better voting" into testable, reproducible computational experiments.

One novelty is the integration of three strands into a single experimental workflow: formal voting theory, behavioural agent-based simulation, and quantum implementations of preference aggregation that can, in principle, be run on quantum computers. Our first quantum experiments indicate that some quantum voting protocols remain stable under realistic noise levels, making them viable objects of empirical study.
A second novelty is methodological: we treat the voting rule as a swappable module in the simulation. This enables controlled comparisons between classical rules (e.g. plurality or ranked-choice) and newer intensity-based mechanisms, using the same simulated population and decision context. The project also introduced new modelling concepts inspired by the Danish concept of ‘Samfundsind’ - such as measuring ‘civic influence’ and explicitly representing tensions between self-interest and the common good - to study how collective outcomes and social cohesion emerge over time.

The potential impacts are practical: better evidence for designing participatory processes, and more transparent tools for stress-testing democratic innovations before deployment. To ensure uptake and credibility, the next steps are to scale simulation experiments, validate models with real pilot data, audit systems for bias and strategic manipulation, and document interfaces and data formats so that public authorities and other researchers can reuse the tools.