Why Scaling Deliberation with AI is a Game-Changer

This is the first of three articles appearing here on how artificial intelligence can scale public deliberation processes to include large numbers of participants, and why this is important for future governance.

Deliberative engagement typically relies on small representative samples of 25-50 people to solve problems. Artificial Intelligence (AI) could scale this to thousands, fundamentally changing how we do deliberation. But once we have a "representative sample" of participants, do we really need to add more? Yes, as recent research indicates, size matters. Let us explain why.

Trans Mountain Pipeline Example

The Trans Mountain pipeline expansion aims to increase oil flow from Alberta to the Pacific Ocean. Despite opposition from environmentalists, Indigenous communities, and some stakeholders, public opinion was initially undecided. The government launched public hearings to win support, but a sophisticated opposition campaign eventually turned most “undecideds” against the project. Investors withdrew, forcing the Canadian government to invest C$4 billion to complete it.

Not so long ago, this kind of organised opposition was almost unknown. Governments had the authority to act, and the public tended to defer to them. The rise of interest groups, social media, and interconnected policy issues has changed the dynamic. Today, government initiatives like this one often face highly organised resistance.

 This change in the political culture complicates governing. Cabinets can develop plans, feel they have a mandate to act on them, but still be blocked by stakeholders and communities. The Trans Mountain Pipeline is only one example.

 Deliberative Dialogue

Deliberation responds to this changing environment. It brings together citizens and/or diverse stakeholders to discuss, debate, and resolve public policy issues collaboratively. A popular approach is to create a "deliberative mini-public"—a representative sample of citizens reflecting the population’s views. This process encourages participants to work toward mutual accommodation and consensus.

 In a mini-public, participants mirror the broader community's diversity, ensuring various perspectives are heard. If the group is truly representative, people outside the process should find someone at the table they can identify with, someone who shares their interests and views, thereby fostering inclusion and trust. As a result, the findings from the process are supposed to carry legitimacy with community members, who, in turn, recognise and accept this body’s legitimacy – at least, that is the theory. But does it work?

 A Catch-22 for Mini-Publics

Let’s return to the Trans Mountain case. Instead of conventional “consultations,” suppose the government had launched a deliberative mini-public. Might this have won over the “undecideds?”

We’re skeptical. Recent research shows that for the public to view mini-publics as legitimate, they must believe that government will act on the findings. Unfortunately, governments are reluctant to make this commitment because they in turn feel mini-publics can’t guarantee public support for their findings.

There is thus a Catch-22 at the heart of this approach: Mini-publics need broad public support to get a firm government commitment to act, and governments won’t provide a firm commitment to act without broad public support. Is there a way out of this impasse? Yes, but we need to revisit how deliberation works.

The Role of Ownership

Deliberation is about finding consensus through a series of trade-offs and accommodations involving all parties. However, there is a profound difference between being an "insider" and an "outsider" in this process. Deliberation empowers participants, giving them a role in decisions and fostering a sense of responsibility and ownership of the outcomes. Participants realign their commitments as they reflect on options, accept trade-offs, and move toward consensus step by step. Deliberation is a genuine learning experience.

 In contrast, outsiders have no direct role in the decisions; they are not part of the step-by-step process, so they don’t realign their commitments or develop the same sense of ownership of the solutions. Indeed, increasingly, they may be part of an alternative participatory process organized by the initiative’s opponents. That’s what happened in the Trans Mountain case – and it’s a growing concern for governments.

Opposition movements today recognise that successful resistance is about size, participation, and ownership, not representative samples. This is the dynamic ground on which public policy battles are fought, and on which public legitimacy rises and falls. While we are not opposed to mini-publics, we believe that high-stakes issues with organised oppositions should be met with large-scale deliberation aimed at fostering broad ownership of the solutions. Size matters. But is this achievable?

Using AI to Scale Deliberation

Scaling participation across a community, a region, or even a country poses a particular challenge: As the number of participants increases, so does the volume of qualitative and quantitative data generated. Using conventional engagement tools to collect, analyse, and manage this data quickly becomes too labor-intensive, time consuming, and costly to be effective, which is why deliberative processes typically are small.

However, as we’ll see in Article 2, artificial intelligence can help solve this data management problem, making genuine deliberation possible in large processes. This in turn fosters ownership among a larger share of the public, giving governments a reason to support these processes. And that could be a game-changer.

Stay tuned.

This article was written by Don Lenihan PhD, Public Engagement Expert, President and CEO of Middle Ground Policy Research, Damian Carmichael former public servant with the Australian Government and Engagement Lead at Converlens and Tom Workman Director Converlens

References

For a discussion on the appropriate size of deliberative models, https://www.publicdeliberation.net/how-representative-is-it-really-a-correspondence-on-sortition/

“The Rise, Fall and Re-Rise of Deliberative Democracy in New Zealand,” by Simon Wright, Tatjana Buklijas, and Max Rashbrooke, in Policy Quarterly – Volume 20, Issue 2 – May 2024.

“Can the use of minipublics backfire?”, by Lisa Van Dijk and Jonas Lefevere https://pmc.ncbi.nlm.nih.gov/articles/PMC10084071/; and “Mini-publics and Public Opinion”, by Shelley Boulianne https://journals.sagepub.com/doi/10.1177/0032321717723507.