Today we restart our Monday series on Governing with AI

Learning the How-To's of AI-Enhanced Public Engagement

Sign up for the Reboot Democracy Workshop Series - Register for individual workshops or the full series here

Abandoned Engagement

Singapore Together launched in 2019 with ambitious plans to make public consultation a regular part of how the government in Singapore would work. The government organized deliberative mini-publics each with fifty participants. Residents discussed pressing issues like recycling and public health. But after just four such exercises, the government quietly shifted away from political participation toward neighborhood volunteering. 

It’s no wonder. The UK Cabinet Office estimates that “a consultation attracting 30,000 responses requires a team of around 25 analysts for 3 months to analyse the data and write the report.”

The challenge wasn't that citizens couldn't engage meaningfully in policymaking—it was that meaningful participation proved too time consuming for the government to organize. 

Even when citizen engagement works, it remains episodic rather than systematic—one-off events rather than integrated institutional practice. Ireland's citizens' assemblies produced major policy wins on abortion and same-sex marriage but happened just six times over fifteen years.

It’s no wonder. The UK Cabinet Office estimates that “a consultation attracting 30,000 responses requires a team of around 25 analysts for 3 months to analyse the data and write the report.”

But now leaders may be able to use AI to reduce the time and cost and, more importantly, go beyond asking to listening to what people have to say

In an effort to increase the regularity with which institutions tap the collective intelligence of our communities, we're kicking off the Reboot Democracy Workshop Series this week with the first of 11+ practical workshops designed to make participation efficient and effective.

In an effort to increase the regularity with which institutions tap the collective intelligence of our communities, we're kicking off the Reboot Democracy Workshop Series this week with the first of 11+ practical workshops designed to make participation efficient and effective.

Together with Professor Danielle Allen and the Allen Lab at Harvard and in partnership with InnovateUS, we will learn together how to use AI to design, run and measure impactful public engagement with AI. 

Our goal: make engagement robust and routine.

Persistent Barriers

At the start of the Obama Administration, we solicited public input on the President’s 100 day agenda. The public shared 44,000 ideas and recommendations.

While there might have been excellent suggestions in the Citizens' Briefing Book, it wasn’t possible to read all of them, so we read none of them. Instead, they were printed, and White House advisor Valerie Jarrett ostentatiously handed the impressive stack to President Obama in a common instance of televised “democracy theatre.”

While there might have been excellent suggestions in the Citizens' Briefing Book, it wasn’t possible to read all of them, so we read none of them.

Having worked in government at the federal and state level in the United States and advised many international government agencies, I can attest that among well-intentioned government leaders (yes, I know that's not all of them), there is a genuine desire to engage.

Public professionals want to learn from those they serve but two challenges are particular impediments:

• First, leaders don't know where to begin: Which tools should we use? What timeline is realistic? How many people do we need with what skills? What will this cost? 

• Second, even when institutions take the plunge, the process proves time-consuming and expensive. It's just too hard and, in the end, so much effort goes into running the engagement that there is no time to act on what is learned and participation remains divorced from impact.

The Reboot Workshop Series: The How-To’s of Participation

The Workshop Series tackles both challenges head-on. 

The 11 workshops are designed to teach the How-To’s of engagement that connects to real institutional decision-making.

• September 11: Why Public Engagement?

• September 24: Different AI Horses for Different Courses: Matching Tools to Purpose

• September 30: Designing Smarter Engagement with AI

• October 14: Solving Problems with the Public

• October 15: Defining Problems with the Public

• October 29: Citizen Assemblies and AI

• November 12: Co-Creating with Multi-Lingual Communities

• November 19: Listening to the Public with AI

• November 20: Inviting, Summarizing, and Responding to Public Comments

• TBD: Deliberating with the Public (with Audrey Tang)

• TBD: Measuring Impact of AI-Enhanced Engagement

We are replacing our usual lecture series with leading thinkers with hands-on workshop series with experienced doers.

Danielle Allen and I, together with our colleagues Sarah Hubbard and Agueda Quiroga, will moderate.

Our goal? Learn how AI can make high-quality engagement more accessible, affordable, and actionable. Develop the practical skills needed to design, implement, and sustain AI-enhanced citizen engagement

“Combinatorial Democracy”

For the public and public servants alike, engagement offers the capacity to reach communities that are often left out, inform how we make decisions from people’s expertise and lived experience, and translate that feedback into decisions that matter, directly improving the quality of governance and the strength of our democracy. 

Introducing AI to help with drafting, summarization and synthesis goes beyond efficiency and even scale. AI offers the opportunity to practice what I call “combinatorial democracy”--the ability to combine multiple styles and forms of public engagement.

AI offers the opportunity to practice what I call “combinatorial democracy”--the ability to combine multiple styles and forms of public engagement.

The way we solve problems in practice is a multi-stage process. We have to define the problem, develop solutions, and create a blueprint to implement those solutions. Yet because of cost and complexity, engagement is too often a one-shot deal with a single platform or method. 

I'm excited that we can now explore how to knit together multiple forms of engagement to understand problems and their root causes, to co-create solutions and their implementation strategies, and to engage people in measuring what works. 

We can combine different modalities and types of engagement:

• We can invite written comments and video responses 

• We can do both representative and non-representative forms of engagement

• We can deliberate and co-design

• We can tap collective intelligence to understand problems and co-create solutions

• We can do short, one-question surveys and multi-day deliberative exercises

• We can combine online and in person forms of participation.

• We can ask kids and adults, public sector and private

In short, when AI brings down the cost and complexity of participation, we can knit together multiple kinds of engagement with different communities across the full spectrum of democratic participation or what I nickname “combinatorial democracy.”

Learning from Doers

This series builds on earlier online courses the GovLab has offered to community leaders, government leaders, and lawmakers as well as shorter workshops in the how-to’s of public engagement. We’ve written up some of those instructions in this Guide the GovLab and Nesta published to help institutions know “how to use new technology to collaborate with the public.” But these resources predate modern LLMs.

Instead, this series brings together leading practitioners using AI to do participation better. Representatives from cities like Copenhagen and Hamburg will show how they've systematically integrated citizen engagement into ongoing governance processes rather than treating it as special events. Beth Goldberg from Google's Jigsaw will share lessons from using AI to process public input in Bowling Green, Kentucky. Audrey Tang, Taiwan's former Minister of Digital Affairs, will demonstrate how AI can support structured public deliberation at scale. 

Making Participation Real, Making Participation Routine

The goal is to make citizen engagement common institutional practice rather than exceptional experiments. When institutions have the practical knowledge and AI-enhanced tools to engage citizens efficiently and effectively, democratic participation can become as routine as budget planning or performance evaluation. That's the transformation we're working toward: moving from episodic engagement to systematic democratic governance where citizen voices genuinely shape institutional decisions.

Register for individual workshops or the full series here