Future Building Starts With Coordination: Broad Listening at ETHPrague

At ETHPrague, Yuting Jiang presented Agora Citizen Network as an open-source deliberation tool for communities facing complex and divisive questions.

The core claim is simple: the hardest problem is not intelligence. It is coordination. From climate change to pandemic response to AI governance, many of our biggest challenges are not caused by a lack of ideas. They are coordination failures.

This talk introduces Broad Listening as a way to move from shared attention to shared decisions.

Key Themes

• Broad Listening: hearing from many people, not only the loudest voices.
• Active Listening: inviting people to participate directly in structured deliberation.
• Passive Listening: analyzing existing discourse from forums, assemblies, meetings, or online conversations.
• Opinion mapping: using agreement and disagreement patterns to reveal groups, tensions, and shared priorities.
• Plural Voting: turning consensus across differences into an actionable ranked list of priorities.
• Decentralized Deliberation Standard: an initiative building an ecosystem of interoperable civic tech tools sharing verifiable collective intelligence data.

Watch on YouTube

Edited Transcript

Hi everyone. Thank you so much for being here for the last session of ETHPrague. My name is Yuting. I am the founder of Agora Citizen Network.

Agora is an open-source deliberation tool that helps communities map disagreement and find consensus on complex and potentially divisive topics. In other words, we try to build solutions that help communities coordinate better in an age of accelerating technologies and growing fragmentation.

I want to begin with a simple claim: the hardest problem in the world is not intelligence. It is coordination.

From climate change to pandemics to AI governance, many of our challenges are not due to a lack of ideas. They are coordination failures.

Someone from Anthropic joked that even with AGI, it would still be impossible to get six people in a room to agree. I think that joke points to something real: even with a shared vision, even with powerful intelligence, we still cannot automatically arrive at shared decisions.

So the question is: how do we go from shared attention to shared decisions?

This is the shift that I want to talk about today: from broadcasting to Broad Listening.

Historically, broadcasting has been one of the most powerful ways to coordinate. From newspapers to radio to television, broadcasting has been weaponized in the past and is highly controlled for a reason.

Social media platforms changed the game and allowed many more people to speak up. Many grassroots movements would not have happened without them.

However, many social media platforms, whether Twitter or TikTok, follow the logic of influencers and followers, which is essentially still broadcasting.

What I am really excited about is Broad Listening: a concept that allows us to hear from millions instead of only hearing from the loudest.

There are two types of Broad Listening tools that we have today.

One is Active Listening tools, which invite active participation from people.

The other is Passive Listening tools, which analyze existing discourse, whether from existing forums or from conversations that are already happening.

There are pros and cons, and there are tradeoffs. Ideally, we should combine both to ensure good quality and good quantity of data.

A famous example of Broad Listening was in Taiwan, where the government invited 4,000 people to participate in an open consultation using pol.is on how to regulate Uber. Citizens, taxi drivers, Uber drivers, government officials, and others deliberated together on the platform, and the results of the deliberation were eventually turned into legislation.

Last year, the city of Tokyo also leveraged Broad Listening in formulating its 2050 strategy.

Inspired by, and learning from, many of these initiatives, my co-founder and I started building Agora, originally funded by Horizon Europe.

The concept is very simple. On Agora, users can create or participate in conversations by clicking agree, disagree, or unsure on each other's opinions. They can also submit their own opinions.

Then we use machine learning to group participants based on how they vote. This step is purely deterministic, which means it does not depend on what language people use. It is based on patterns of agreement and disagreement.

The reason we do this first is to make the outcomes of the opinion mapping reproducible and auditable. After that, we use an LLM to provide short labels and summaries for these groups, so people can make better sense of them.

What is really interesting about this approach is that we are moving from majority vote to consensus across differences.

A lot of the time, what is most interesting is not the majority opinion. It is the bridging statement: the statement that multiple groups share, even between a small minority group and a large majority group, although they may disagree on many other things.

Once we identify these consensus points, we can invite people to prioritize them and express how much they agree with a list of proposals, in a step we call Plural Voting.

We take into account demographic data and opinion group data to build collusion resistance. This step is important because it produces a shared ranked list of priorities that is immediately actionable for a group.

Agora is based on decentralized identifiers, or DIDs. It is fully compatible with zero-knowledge proofs, or ZKPs. For now, we are compatible with Zupass and ZK Passport developed by the Rarimo team behind the Freedom Tool.

This allows people to prove citizenship, age group, and other attributes anonymously, without sharing personal details. Agora is also fully open source.

To make this a little bit more concrete, we are currently working with a group of facilitators and sociologists to carry out a local development project in the south of France.

In this consultation, the sociologists are carrying out qualitative interviews with representative stakeholders: farmers, union representatives, civil servants, and others. They are identifying points of tension and trying to understand the dynamics around water use and water management.

The insights from these interviews will be used to seed an open consultation on Agora, which can then scale to many more stakeholders.

This is just one example of a Broad Listening process. We could also imagine mixing many other steps and tools. For example, we could have an in-person local assembly where we also use a Passive Listening tool to listen to what is happening and synthesize the insights.

I say this because I think how we design a process is as important as, if not more important than, the tools themselves.

This is also why we launched the DDS initiative to build a Decentralized Deliberation Standard that enables an ecosystem of interoperable collective intelligence tools.

These tools could include dialogue tools, voting solutions, decentralized social apps, or even prediction markets. Using interoperable and verifiable data, communities, institutions, DAOs, network states, and companies can make better collective decisions.

This is also why I think this conversation belongs at ETHPrague, where we are actively trying to make sense of realities that are plural, and to build better futures.

I want to end the lightning talk with one last message: we do not just predict the future. We co-create it.

Thank you so much.

Keywords: Agora Citizen Network, ETHPrague, Broad Listening, Active Listening, Passive Listening, collective intelligence, coordination failures, digital deliberation, opinion mapping, consensus across differences, Plural Voting, decentralized deliberation standard, DDS, zero-knowledge proofs, decentralized identifiers, Zupass, ZK Passport, civic technology, DAOs, network states, AI governance.