TECHNOLOGY PARTNERS

Build replaceable components — and evidence that outlives the vendor.

ETHICA’s technical goal is not a central “democracy platform”. It is a set of federated components with explicit security boundaries, open interfaces, version history, reproducible assurance and human review paths.

Build a knowledge commons — not a digital oracle.

Citizens meet one coherent portal. Behind it, identity, deliberation, evidence, rule simulation, public authority and audit remain separate trust domains. A language-model error must not become a population-register error merely because both live in the same presentation slide.

01FactsWhat do the sources say?
02ModelHow is the information interpreted?
03ValuesWhich legitimate goals conflict?
04MandateWho may choose and carry responsibility?HUMAN DECISION
  1. 1Human encounterPlain language, accessibility, language support, proxies and visible human help.
  2. 2Identity & consentSelective disclosure, data minimisation and separate evidence chains.
  3. 3DeliberationProposals, objections, opinion maps and minority voices.
  4. 4EvidenceTraceable sources, knowledge graphs, model cards and uncertainty.
  5. 5Rules & scenariosExecutable rules, budget tests, distribution and consequence cards.
  6. 6DeliveryApproved actions, named responsibility and observed outcomes.
CONTROL SPINEVersion log · source provenance · permissions · incidents · independent audit · appeal · fallback

Much of the infrastructure already exists. The missing piece is the coherent mandate chain.

ETHICA need not reinvent identity, data exchange, participation and rule simulation on the same Monday. Reuse reduces risk and focuses development on rights gates, measurement contracts and reconsideration.

X-Road / X-teeSigned data exchange among independently controlled systems.
GOV.UK One LoginOne entry point without turning every service into one system.
EU Identity WalletSelective disclosure and open reference components.
DecidimInitiatives, collaborative drafting, participatory budgets and result reporting.
PolisLarge-scale opinion mapping that shows consensus and division.
OpenFiscaExecutable and comparable tax and benefit rules.
2026–2031 · HIGH PROBABILITY

Better multilingual assistants for text, speech, images, documents and code. They compare rules, detect contradictions and perform bounded workflows with human approval at high-risk steps.

2032–2038 · MEDIUM-HIGH PROBABILITY

Continuously updated knowledge graphs, competing analyses, stronger causal policy tests and more verifiable identity, delegation and logging.

2039–2046 · LOW-MEDIUM PROBABILITY

More dynamic societal simulations and stronger research support — but no evidence-based promise of legitimate automated government.

Identity without vote disclosure

Eligibility and ballot secrecy must live in separate evidence chains.

Verifiable lottery

Population, randomness source, algorithm and result must be independently reproducible.

Evidence & model gateway

Separate source retrieval, reasoning, rule checks and explanation. Version everything.

Federated service domains

Minimum necessary data, bounded scopes and no central raw-data database.

Assurance by design

Connect constitutional norms to requirements, code, tests, release and observed outcomes.

Fallback & reversibility

Manual fallbacks, rollback and multiple independent emergency brakes.