Contribution to the handbook
Priority · Assessed 2026-10-28
The distinction between preference aggregation and approaches such as consensus is material to considering alternative ways citizens might make group decisions. The study’s limits identify an important uncertainty, not proof that any alternative is better.
Does this challenge the goal?
In-scope gap
The question identifies untested forms of group decision-making within the existing goal; it does not contest that goal or its beneficiary boundary.
The supplied evidence identifies preference aggregation and consensus as distinct areas, but does not establish how the untested approaches would work or perform.
Why this category?
- Identifies meaningful alternatives to preference aggregation, including consensus approaches.
- Highlights a consequential limitation in evidence about group decision-making; the supplied materials do not show whether AI would improve or harm these alternatives.
The evidence comes from one study’s stated scope and limitations; it does not establish that every other group-decision method is untested or how any method would perform in governance.
This is an AI assessment of contribution to the goal, separate from whether the question has an answer.
Making and interpreting laws →
Tests, stop conditions and unresolved work →
Read the exploration goalIdentify the group decision-making forms not tested and distinguish them from the preference-aggregation task studied.
What we can say so far
The study left **consensus-building** untested: it did not test whether AI agents could help people reach agreement through multi-agent consensus, or how conventional consensus methods might change when AI agents represent preferences. The authors distinguish these questions from the study’s preference-aggregation task, which used pairwise policy choices to estimate or reproduce group preferences. The study also did not use multi-agent AI systems; cited findings about their possible accuracy come from other research, not this study.
What remains open
- The supplied material does not report whether AI agents improve or harm consensus.
- It identifies these untested areas but does not provide an exhaustive inventory of every possible group decision-making form.
Conditions and limits
- This is the authors’ stated scope and research limits as reported in the supplied material, not an independent assessment of the full study.
- The cited possible accuracy of multi-agent systems comes from other research and is not evidence about consensus or an effect demonstrated by this study.
Evidence and related ideas
Supporting evidence, qualifications and context play different roles in this answer.
context
The volunteer policy data The study uses volunteer choices between randomly paired policies from the two leading candidates’ 2022 Brazilian presidential programs. The choices came from an experiment involving collaborative construction of a government program.
Describes the study’s paired-policy-choice data, helping distinguish its task from consensus-building.
Large Language Models (LLMs) as Agents for Augmented Democracy
Read full source: Large Language Models (LLMs) as Agents for Augmented Democracy (opens in a new tab) ↗Read source discussion: Large Language Models (LLMs) as Agents for Augmented Democracy →
context
Estimates of group preferences Using cross-validation, the authors report that a probability-based sample augmented with model predictions estimated group policy preferences more accurately than the sample alone.