AI can help humans find common ground in democratic deliberation
The study this post is about. Full text is paywalled, but the abstract is free — and the link is still worth clicking if you have institutional access or are willing to pay. DOD's write-up below draws on the secondary sources listed at the bottom for everyone else.
Finding agreement through a free exchange of views is often difficult. Collective deliberation can be slow, difficult to scale, and unequally attentive to different voices. In this study, we trained an artificial intelligence (AI) to mediate human deliberation. Using participants’ personal opinions and critiques, the AI mediator iteratively generates and refines statements that express common ground among the group on social or political issues. Participants (N = 5734) preferred AI-generated statements to those written by human mediators, rating them as more informative, clear, and unbiased. Discussants often updated their views after the deliberation, converging on a shared perspective. Text embeddings revealed that successful group statements incorporated dissenting voices while respecting the majority position. These findings were replicated in a virtual citizens’ assembly involving a demographically representative sample of the UK population.
Someone in DOD flagged this one around and the reaction has been consistently "wait, this is actually interesting" — so here's the heads-up version. Google DeepMind built an AI system, named after Jürgen Habermas, whose only job is to write a group statement that a divided room can agree describes their discussion fairly. In controlled experiments with UK citizens, it beat human mediators at that job more often than not — published in Science in October 2024, with a follow-up essay from the same team a year later if you want the caveats too.1
Loomio, DemocracyOS, Pol.is, Decide Madrid, Decidim, Democracy Earth, RadicalxChange — the Democracy Landscape already documents all of these individually. What it hasn't done yet is put them next to each other. Read as a sequence rather than a list, they tell one continuous story: a protest movement's decision-making problem, handed to software, again and again, for fifteen years.
The DOD podcast archive runs from 2017 to 2023. DOD is a discussion forum — these were working conversations among practitioners thinking through how democracy functions and fails, not policy proposals or predictions. Revisiting them from 2026 is an exercise in calibration: did the observations hold up? Where did the analysis prove accurate, where is it still unresolved, and what did it miss entirely?
This is an external event organised by Beyond Billionaires — not a DOD event. We're flagging it because the topic and speakers are directly relevant to our work, and DOD member Craig Lambie is one of the speakers.
Taiwan is frequently cited as a place where digital deliberation actually worked at government scale. That claim deserves scrutiny — and the scrutiny is more interesting than the headline. The factual background is in the vTaiwan and g0v entries in the DOD Democracy Landscape.
This event in victoria is of interest to us so passing around the telegram chat and to various other Designing Open Democracy members in Victoria, because it's about open governance but via Citizen Assemblies.