💡 In simple words: Imagine a politician who signs a fair promise before an election: to work on what most people really need, instead of fighting left against right. A computer helper (AI) checks if the promise is kept, and a group of humans checks the computer helper. If the politician breaks the promise, money they set aside is taken away. This idea is only at its very beginning.

🎯 In 20 seconds (scientific summary): This candidate pilot explores a contract framework for political candidates. A candidate would formally commit to needs-based positions, grounded in statistics and independent of the left-right divide, to respect popular votes, and to carry initiatives, new legislation and amendments aligned with the commons and the majority will as clarified by honest polling. Compliance would be verified by AI tooling supervised by a human verification committee, and backed by a financial guarantee with compensatory deductions in case of breach. The agreement is intended to accompany the candidate through every stage: support committee, candidacy, campaign, election and mandate. This page is under construction: it lays out first foundations only, as a basis for discussion.

Pilot: AI-Driven Politician Agreement

⚠️ This page is under construction. It lays out the first foundations of the concept. Nothing here is decided or operational: it is a candidate pilot under study, proposed as a basis for discussion.

The concept

The AI-Driven Politician Agreement is envisioned as a contract between a political candidate and the citizens who support them. By signing it, the candidate would demonstrate a formal commitment (not just a motivation) to:

defend needs that are independent of left and right: balancing needs identified through statistics, not through the left-right divide;
respect popular votes and get personally involved in making them happen;
carry initiatives, new articles of law, amendments and modifications that are always based on the common good and on the popular will of the majority;
rely on the clarity of the majority will: on most subjects, the popular will is not hesitant, it is clear. This clarity comes from polls built on good, honest questions, not on questions that project fears.

AI verification and the human verification committee

The founding principle of the agreement is that the candidate's message and the respect of their commitments would be verified by AI. This requires a clear introductory rule: behind the AI, there are humans.

A verification committee would stand behind the AI tooling. This committee:

checks that the prompts used are correct and fair;
chooses the AIs and makes their usage evolve, transferring from one AI to another as technology evolves;
guarantees at the same time quality, honesty, ethics, audit and transparency;
maintains the indicators measuring whether the politician respects the commitments they have taken.

This intends to follow the same logic as WikiDeal's broader approach to internal AI use: AI as a tool, humans as the accountable layer.

The agreement at every stage

The contract is intended to accompany the politician at every step of a candidacy.

1. Creating a support committee

The first step would be the creation of a support committee, following the standard practice of electoral candidacies at any level (communal, cantonal or regional, national). Comparable standard forms include campaign committees and campaign organisations, political action committees in the United States, and the "comité de soutien" tradition in French-speaking countries.

A guiding principle: the higher the level of the election, the larger the support committee must be. Otherwise it remains a support committee of principle only.

The committee is intended to evolve over time. Anyone could become a member, under conditions defined by each support committee and each politician, provided those conditions remain ethical, honest, acceptable, open, and based on the majority and on real needs.

2. Filing the candidacy

The formal deposit of the candidacy, accompanied by the signed agreement.

3. Campaign

During the campaign, the candidate would build a program based on the most common needs, as identified by honest polling and statistics, with AI-verified messaging supervised by the verification committee.

4. Election and mandate

Once elected, the commitments keep applying: involvement, votes and legislative work would be tracked against the criteria defined in the agreement.

Compensatory measures

The agreement would include financial accountability:

the politician deposits a guarantee sum backing the respect of their commitments, in a logic comparable to an escrow held by a notary;
the expected votes and involvements are written down in black and white in the contract;
if the politician does not respect their obligations (to get involved, and to vote in a defined way), compensatory deductions would be applied to the guarantee;
the criteria are defined by the citizens who joined the support committee, and evolve with the committee itself.

Domain examples

The concept applies to any policy domain. Two illustrative examples follow.

Transport

Needs-based commitments in transport could include: free public transport, a surtax on non-shared private transport, incentives against driving alone (refusing carpooling or hitchhikers when sharing is possible), carpooling applications that effectively tax solo driving, and a principle of transport sobriety: lower maximum speeds and real measures to pollute less and to help each other in solidarity.

Housing

Needs-based commitments in housing could include: public-utility housing, a strong and better applied pre-emption right, and incentive measures such as loans allowing housing cooperatives to act upstream of building sales: direct offers to the seller, and early proof that they hold the equity needed to compete against private buyers. The goal: exiting real-estate speculation.

Gov/en/Portal:R&D/Pilot-AI-Driven-Politician-Agreement