Workshop Overview

Local and federal agencies are increasingly using AI to improve public services, supporting decisions in urban planning, security, infrastructure, and essential resource distribution. As the closest tier to citizens, local governments must uphold democratic values and trust, especially in AI-driven smart city initiatives. Yet traditional participatory methods struggle to meet the higher standards required for public sector AI. This workshop will examine emerging practices in participatory algorithm design, focusing on community engagement and responsible adoption.

Workshop Sessions

    Session 1 (90 minutes): Methods and Measures8:00 - 9:30 PM ET

    • PD research often lacks ways to show that participation produced meaningful design or policy changes. While the NIST AI RMF outlines “govern, map, measure, manage” actions, turning these into concrete, measurable practices is difficult. In this workshop, participants will review diverse PD methods and case studies to propose measures that work across approaches. We will start with practical, low‑burden metrics: (a) input-to-decision traceability via change logs linking public input to outcomes, (b) PD coverage across lifecycle stages, and (c) participation breadth and accessibility (who could join, and what language or compensation supports existed). Building from these scaffolds, attendees will identify high‑value measures, compare easy vs. hard-to-collect options, and discuss opportunities and risks in operationalizing both.
    • Through this process, participants will co-design candidate measures and their associated trade-offs to create more concrete tools for evidencing participation in public-sector AI.

    Session 2 (90 minutes): Procurement and Contracts 2:30 - 3:30 AM ET

    • Participants will complete a structured, hands-on exercise on embedding community-engagement requirements into vendor agreements. In small groups, they will review sample procurement contracts and use affinity diagramming to identify where participatory language could be added or strengthened. Groups will then target specific components—like evaluation criteria, vendor reporting, and milestone deliverables—and draft candidate requirements such as engagement plans (including which communities are involved and compensation) and influence logs linking stakeholder input to design changes. Ideas will be grounded in precedents including Amsterdam’s AI Procurement Strategy and Sloane et al.’s AI and Procurement primer, which highlight transparency, engagement, and the tradeoffs public buyers face.

Workshop Themes

    Our workshop focuses on three objectives alongside case studies showcasing participatory AI in public innovation. First, we will examine methods used in public initiatives, assessing strengths and weaknesses for community engagement. Next, we’ll discuss frameworks to enhance involvement and measure outcomes. Finally, we’ll address procurement challenges and the importance of participatory requirements in contracts.

    Theme #1: Evaluating Participatory Methods: Strengths, Weaknesses, and Impact on Community Engagement

    We will evaluate participatory methodologies currently used in public sector AI initiatives to assess how effectively these methods facilitate community engagement. This involves examining approaches such as citizen panels, co-design workshops, and public consultations. We will identify the strengths that enable successful collaboration, such as transparent communication and inclusive representation, and weaknesses, like lack of follow-up or insufficient diversity in stakeholder input, that may limit the impact and trustworthiness of these processes.

    Theme #2: Defining Key Elements and Outcomes for Meaningful Public Involvement in Public Sector AI

    Next, we will focus on conceptualizing essential elements that can foster deeper community involvement in public AI initiatives as well as determining clear, measurable outcomes to assess the success of public engagement efforts. We will explore key factors that make participatory processes meaningful and impactful, such as building trust, ensuring transparency, fostering inclusivity, and creating opportunities for genuine influence on decision-making.

    Theme #3: Tackling Procurement: Embedding Participation Requirements in AI Vendor Contracts

    Finally, we will explore how public agencies can integrate participation requirements into contracts, especially considering that most agencies do not have the capacity to build AI systems in-house and often procure AI solutions from private vendors. It is essential that these vendors follow principles of community engagement, transparency, and accountability. We will discuss how specific participation guidelines in procurement contracts can be embedded that clearly outline expectations for vendors to involve communities in the design, testing, and evaluation of AI systems. This can include contractual clauses mandating public consultations or co-design sessions, such as requiring vendors to hold community workshops to gather citizen input and ensure diverse stakeholder representation in decision-making.

Contact: participatoryaidesign@gmail.com

Cover image: PACIFICO Yokohama