PolicyCraft: Supporting Collaborative and Participatory Policy Design through Case-Grounded Deliberation

TL;DR

PolicyCraft introduces a case-grounded deliberation approach to collaborative policy design, enabling communities to propose, critique, and revise policies through concrete cases, discussion, and voting. In field studies across two university courses, policies developed with PolicyCraft achieved stronger support and greater consensus than a baseline system, with participants better understanding one another’s positions and more frequently iterating based on shared cases. The work demonstrates the value of grounding policy language in real scenarios to improve legitimacy, while also highlighting challenges in balancing specificity with generality and in synthesizing a coherent final policy set. It further discusses opportunities to scale case-grounded design, enhance sensemaking, and advance AI-assisted tools within participatory governance across diverse contexts.

Abstract

Community and organizational policies are typically designed in a top-down, centralized fashion, with limited input from impacted stakeholders. This can result in policies that are misaligned with community needs or perceived as illegitimate. How can we support more collaborative, participatory approaches to policy design? In this paper, we present PolicyCraft, a system that structures collaborative policy design through case-grounded deliberation. Building on past research that highlights the value of concrete cases in establishing common ground, PolicyCraft supports users in collaboratively proposing, critiquing, and revising policies through discussion and voting on cases. A field study across two university courses showed that students using PolicyCraft reached greater consensus and developed better-supported course policies, compared with those using a baseline system that did not scaffold their use of concrete cases. Reflecting on our findings, we discuss opportunities for future HCI systems to help groups more effectively bridge between abstract policies and concrete cases.

Figures (10)

• Figure 1: PolicyCraft's overall process.
• Figure 2: PolicyCraft's "policy page" for a given policy. On the left side is the navigation bar, where users can click to visit the Policy Repository to see all current policies, and the Case Repository to see all cases. Users can also visit the About page to view information about the current policy development campaign and to discuss general topics, not specific to a particular policy or case. In the center of the screen, the title and description of the selected policy is shown, along with its related cases. The visualization of users' votes on the cards is inspired by the design of Polis small2021polis. The bar represents how many people have voted out of the total number of users, while the percentages reflect the distribution of votes—whether to allow, disallow, or indicate unsure—among those who have already voted. Users may add new related cases by clicking the "Edit cases" button, which brings them to the page shown in Figure \ref{['fig:edit_case_page']}. They can also revise the policy by clicking the "Edit policy" button, which brings them to the page shown in Figure \ref{['fig:edit_policy_page']}. Users can also start a discussion thread about this policy and close it once the topic has been resolved. Finally, users can propose new policies by clicking the "Create" button in the top-right corner.
• Figure 3: The workflow for editing a policy's "related cases" in PolicyCraft. The upper half of the screen shows the cases currently associated with a policy. The bottom section allows users to add additional cases either by searching the case repository with keywords (e.g., "brainstorm," as shown in this figure) or by authoring a new case. When adding a case to a policy, users label it to indicate whether they believe the current version of the policy would allow it, disallow it, or whether it is ambiguous how the policy would treat this case. Users can optionally brainstorm with the built-in AI assistant to generate cases that illustrate the policy or identify its potential flaws.
• Figure 4: (a) Once a user adds the case shown in Figure \ref{['fig:edit_case_page']} to the policy, it will appear as a card in the related cases section of the policy. The yellow message highlights the misalignment between the label linking the case to the policy (allow) and the majority vote on the case (disallow). (b) Users can click on the card to open a modal with additional details about the case, including other users' reasons for wanting to allow or disallow it. (c) For more in-depth discussions, users can click the "open case" button to visit the case page.
• Figure 5: The workflow for editing a policy in PolicyCraft. Users first update the policy's wording and then check if the edit requires updating the labels for any of the related cases. Users may also brainstorm with the built-in AI assistant to improve the policy. For example, the policy description shown in this figure is suggested by the AI assistant based on the user's interactions shown in the right-side panel.