Hearing how much people care, not just how loud they shout
Most public consultations count how many people said yes, and let the most persistent commenters, and increasingly bots, dominate. Two techniques do better. Giving every verified resident the same fixed budget of votes shows how strongly people care, and summarising views as a slate that represents every sizeable group shows where they agree.
What goes wrong with polls and comment boxes
A yes-or-no checklist treats every vote the same. A neighbour who mildly wants a repainted lobby counts the same as a resident who needs a step-free entrance to reach the street. The cosmetic project wins because more people walk past it.
Open comment periods reward volume. One person with free evenings, or one script, can file fifty submissions. Officials then read the loudest and most numerous voices, not the most affected ones.
Summaries erase minorities. Whoever condenses two hours of discussion into five bullet points decides which views survive. An automated summary that aims for the "average" view can describe a position nobody in the room actually holds.
The alternative many people reach for, letting people pay to back what they care about, measures wealth along with care.
Measuring how much people care
Give every verified resident the same budget of votes to spread across the items on the agenda. Make each extra vote on the same item cost more: one vote costs one credit, two cost four, three cost nine. This is quadratic voting.1 Everyone starts equal, so money can't buy more voice. A small group whose daily lives depend on one item can concentrate their budget and be seen. And because piling votes on a single item gets expensive quickly, no organised faction can sweep every item at once.
Finding what people agree on
Instead of one summary, produce a short slate of statements chosen so that every sizeable group of participants sees its view represented. Researchers have shown how to do this with language models while keeping mathematical guarantees of proportional representation, and have tested it on opinions about city improvement measures.2,3 Bridging tools such as Polis add statements that people from opposing clusters both endorse, which is where common ground actually is.4
What a consultation would look like
- Residents prove they live in the area without handing over ID, using a credential that shows only that one fact (see proving one fact without handing over the rest). A flood of bot submissions then counts for nothing.
- Participants discuss the options first, with facts presented separately from opinions. Deliberation before counting is the core of the citizens' assemblies and mini-publics many governments now run.5
- Each resident spreads a fixed budget of votes across the items.
- Officials receive the vote totals, the intensity behind them, and a proportional slate of the arguments made, including the minority ones.
The objection
Quadratic voting is easy to game if one person can vote under several identities, or if people pool their budgets. That is why it only works together with reliable one-person-one-budget checks, and why results should be published so collusion can be spotted. No voting rule escapes trade-offs altogether: Arrow showed that no ranking method satisfies every fairness condition at once.6
What's unsolved
- Proving someone is a real local resident without building a register that could be used for surveillance.
- Preventing vote-budget pooling and coercion when voting happens on a personal device.
- Checking that a language model's slate of statements represents people faithfully, not just according to its own estimates.
Sources
1. Steven P. Lalley and E. Glen Weyl, "Quadratic Voting: How Mechanism Design Can Radicalize Democracy," AEA Papers and Proceedings 108 (2018): 33–37, doi:10.1257/pandp.20181002.
2. Sara Fish, Paul Gölz, David C. Parkes, Ariel D. Procaccia et al., "Generative Social Choice" (2023), arXiv:2309.01291.
3. Niclas Boehmer, Sara Fish and Ariel D. Procaccia, "Generative Social Choice: The Next Generation" (2025), arXiv:2505.22939.
4. Christopher T. Small, Ivan Vendrov, Esin Durmus et al., "Opportunities and Risks of LLMs for Scalable Deliberation with Polis" (2023), arXiv:2306.11932.
5. OECD, Innovative Citizen Participation and New Democratic Institutions: Catching the Deliberative Wave (OECD Publishing, 2020), doi:10.1787/339306da-en.
6. Kenneth J. Arrow, "A Difficulty in the Concept of Social Welfare," Journal of Political Economy 58, no. 4 (1950): 328–346, doi:10.1086/256963.
