💪🤖 New Paper: Deliberative Muscles & AI
Our latest paper is now live - join us for the launch event at 17:00 CEST today!
Democracy is under strain, and one of the most promising responses to that strain is the growing global movement around deliberative assemblies: citizens’ assemblies, citizens’ juries, and related forums that bring randomly selected, broadly representative groups of people together to weigh evidence, listen to one another, and make shared decisions on complex public issues. Over 1,000 such processes have now been run worldwide, and a growing body of evidence suggests they depolarise opinion, generate well-reasoned recommendations, build trust, and reconnect people to political life.
However, these processes are also resource-intensive, slow, and hard to scale, and have thus become a site of intense interest for AI integration. The pitch from many technologists, practitioners, and funders is consistent: AI can make deliberation cheaper, faster, more accessible, and more scalable.
In this paper, Claudia Chwalisz, Sammy McKinney, Jorim Theuns, and Eugene Yi argue that AI, when designed with care, can indeed play a powerful role in strengthening deliberation. But the very efficiencies that make AI attractive also risk undermining what deliberation is for in the first place.
Their starting point is that deliberative assemblies are not decision-making machines whose sole value lies in the recommendation they produce. They are also spaces in which participants exercise and develop the civic capacities that democratic life depends upon. If we automate too much, we may end up with smoother processes that hollow out the productive friction that makes them valuable, while simultaneously reducing people’s ability to participate in democratic life.
The authors make three contributions:
First, they argue that one of the most important and most overlooked virtues of deliberative assemblies is that they build deliberative muscles: the cognitive, dispositional, and relational capacities that citizens need to do the work of democracy together.
Second, they offer a typology of seven deliberative muscles: self-reflection, reasoning, dialogue, vulnerability, collaboration, imagination, and facilitation. For each muscle, they map AI uses. Context matters throughout: the same tool can strengthen a muscle in one setting and substitute for it in another.
Third, they argue that the exercise of these seven muscles, sustained across many people and many processes over time, is what produces the societal capacity that resilient democratic systems require. This has implications for civic infrastructure: for the training, standards, and ecosystem-building bodies that sustain deliberative practice; for how we think about AI ownership, provenance, and dependency; and for the case for a civic AI future in which communities are not merely end-users of DelibTech but central to the conditions that make it work at all.
The central design question for anyone building, commissioning, or funding deliberative technology is therefore not whether AI produces a recommendation faster, but whether it leaves participants, facilitators, and the wider institutional ecosystem more capable at the end of the process than at the start.
Join the authors, together with respondents Lisa Gutermuth (Mozilla Foundation), Mauricio Mejia (Advisor on democratic change), and Maggie Hughes (CEO and Co-founder, Convoca), for this launch event to discuss the paper and its implications for different sectors. The event will be moderated by Andrew Sorota (Office of Eric Schmidt).
What to expect:
Why deliberative assemblies are sites where we exercise and strengthen the capacities that democratic life requires
The seven deliberative muscles and how technology can help or hinder each
The distinction between complementary and competitive AI uses in deliberative practice
What this means for practitioners deciding how to integrate AI, technologists deciding what to build, and funders deciding what to back.
This paper is for people with the authority and responsibility of deciding how technology is going to be deployed in a deliberative process - commissioning authorities in government and other organisations, as well as democracy practitioners, designers, and facilitators. We are also writing for technologists deciding what to build, for funders deciding what to back, and for scholars mapping a fast-moving field.
On the radar
💻 People Powered are hosting ‘1,000 Citizens’ Assemblies Later, What’s Changed?’, an online event, on 23 July, 17:00 BST. They will share the latest findings and lessons learned from where assemblies have driven real policy change.
📊 Navigating rising expectations and new horizons, the OECD have published the results of their survey on drivers of trust in public institutions 2026.
⭕️ A new Demos report by Miriam Levin examines everyday democracy - a new democratic operating model to rebuild trust between state and citizen.
🎬 ‘Goodbye Politicians, Hello Democracy’, is a new film by Adam Conkright, available to watch for free here.
Events
As our Cities Programme comes to a close, James MacDonald-Nelson and Hannah Terry examine in greater detail the learnings from working with cities and regions to institutionalise citizens’ assemblies - Vilnius (Lithuania), Kerewan (The Gambia), Esch-sur-Alzette (Luxembourg), and Central Oregon (USA).
I've signed up, very interested in this topic, I've successfully used AI facilitated assemblies to bring rooms full of technical experts to deliberate and agree on sustainability tech standards, I can see its power and I agree it's not about getting to a decision (although, that's incredible important) but it's in the facilitation of deliberation, which, is not easy for humans to do, just building the rapport so you can trust someone else with your opinion/ideas, takes a lot of time IMO.