Nahir — Exploring ideas together

Contribution to the handbook

Priority · Assessed 2026-10-28

This asks about a consequential alternative to choosing policy bundles: expressing preferences on individual policies. The evidence describes that as a possible way to support more detailed participation, but also cautions that prediction accuracy does not establish that agents represent people well, that citizens would use or trust them, or that the approach works beyond the study group. Those distinctions matter to design and testing citizen-input mechanisms.

Does this challenge the goal?

In-scope gap

The question explores how citizens might express policy preferences within the existing goal of serving citizens; it does not contest that goal or its beneficiary boundary.

The evidence presents policy-level participation as a proposal, not as a demonstrated or preferred governance arrangement.

Why this category?

Meaningful alternative for expressing citizen preferences
Consequential representation limits and testing needs

The records arise from one shared study source, not independent evidence. The study’s reported prediction results do not establish real-world representation, public acceptance, or effects on binding decisions.

This is an AI assessment of contribution to the goal, separate from whether the question has an answer.

Making and interpreting laws →

Authority, accountability and resilience →

Tests, stop conditions and unresolved work →

Read the exploration goal

Explain the potential shift from expressing a choice among candidate or party bundles to expressing preferences on individual policies, without assuming that agents achieve it in practice.

What we can say so far

The authors propose that software agents could help citizens express preferences on individual policy proposals, rather than having to choose only among the bundles offered by candidates or parties. This would shift the level at which citizens can state their preferences; the proposal does not establish that agents will be used or achieve this in practice.

What remains open

The supplied evidence does not establish whether agents would be adopted or how individual policy preferences would affect real decisions.

Conditions and limits

This is the authors’ proposal as represented in an AI interpretation of the preprint, not independent new evidence.
The bundle description is a benchmark framing and does not establish that every voter chooses this way.
Whether citizens would use or trust agents, and how their input would affect decisions, remains unresolved.

Objections to consider

The bundle framing is a useful comparison, but the evidence does not show that all voters actually choose in this way. The answer describes it as the candidate-or-party benchmark rather than a universal account.
The authors’ claim that bundles may not fit every citizen is not quantified. The answer presents individual-policy expression as a proposed alternative, not as a proven improvement in fit.
The proposal that agents could support detailed participation is not tested as a practical solution. The answer preserves it as a proposal and makes no claim that agents achieve the shift in practice.

Evidence and related ideas

Supporting evidence, qualifications and context play different roles in this answer.

supports

Choose among individual policies Software agents could help citizens express preferences on individual policies instead of choosing only among policy bundles offered by parties or candidates.

Directly states the proposed shift from choosing policy bundles to expressing preferences on individual policies, with practical effects left unresolved.

Large Language Models (LLMs) as Agents for Augmented Democracy

Read full source ↗

← Back to questions