OpenCandidateDownload old PDF

Concept memo

The positive AI outcome is governance that can finally see the whole system.

OpenCandidate is an argument for full embrace. AI is not campaign software here. It is the public reasoning layer for a world that already needs global coordination and has no human institution capable of doing it alone.

Problem

Politics is being asked to coordinate climate, migration, compute, energy, public health, supply chains, money, and war through institutions designed for slower facts and smaller blast radiuses.

The normal answer is more process. Another committee. Another hearing. Another treaty draft. That answer is not serious at the speed and scale of the systems involved.

Global governance requires AI systems. The question is whether those systems are private, hidden, and captured, or public, contestable, and bound to explicit values.

Solution

Turn a candidate into a public AI governance system. Before election day, the candidate publishes a governing pattern: values, priorities, accepted tradeoffs, rejected tradeoffs, fiscal posture, affected publics, and model assumptions.

After that, major decisions pass through the same public reasoning layer. The AI recommends. The human decides. The record shows whether the decision preserved the governing pattern, updated it, or broke it.

Eigenist frame

Eigenism treats identity as a pattern that can be preserved, copied, updated, and degraded over time. That maps cleanly onto politics. A candidate's public identity is a pattern of commitments, choices, explanations, and tradeoffs.

The old question was whether a politician kept a promise. The better question is how much of the published governing pattern survived the decision. That lets the system distinguish continuity, update, override, and drift.

Mechanism

1.

Publish the candidate's governing identity as a machine-readable pattern.

2.

Apply the AI system to concrete decisions: legislation, budgets, land use, emergency response, procurement, climate, and cross-border effects.

3.

Score connectedness between the action and the governing pattern. Discount boilerplate. Reward specific commitments that actually constrain behavior.

4.

Preserve overrides and model changes in a public continuity ledger.

Why this is the optimistic path

The fearful path treats AI governance as something to delay until institutions feel ready. The captured path lets private systems become the invisible constitution. The protopian path builds public AI governance early enough for citizens to see it, contest it, and improve it.

This is wacky in the way the future is often wacky before it becomes infrastructure. Courts, markets, central banks, epidemiology dashboards, and weather forecasts all made governance more model-mediated. AI is the next layer. Hiding from that does not make it less true.

What it looks like in practice

Candidate as public model

The platform becomes a structured pattern of values, priorities, tradeoffs, hard lines, and affected publics. Citizens can interrogate it before and after the election.

Decision simulator

Every major action is run through the AI governance layer before the vote. The model names matched rules, violated rules, and expected continuity impact.

Override ledger

Human officials can break from the model. The break is public, scored, explained, and preserved as part of the governing identity over time.

Why elections first

Elections are where the public grants authority to a governing identity. That identity should be computable before it becomes power.

Start with candidates because they can consent to the system. Start locally because the decisions are concrete. Build toward global coordination because the problems already crossed the border.

What we need

A first candidate willing to say the quiet part out loud: AI is civic infrastructure

Policy and technical advisors who can turn platforms into model-readable governing patterns

Funders who want the positive AI future built in public rather than argued about forever