At the Open Government Partnership Summit in October 2025 Participation AI and Connected by Data will launch a set of principles for public engagement in AI procurement.
In this brown-bag lunch at the OpenGovHub in Washington DC we’ll be sharing the draft principles for discussion, and gathering shared insights and inputs on how they can be put into practice.
How can different publics and civil society play a greater role in the procurement of AI-driven systems in the public sector? How can AI procurement decisions include the voices of the most affected communities?
At the Open Government Partnership Summit in October 2025 Participation AI and Connected by Data will launch a set of principles for public engagement in AI procurement that provide:
- A clear vision for how public participation and stakeholder engagement should shape public sector AI
- Practical steps and guidance for public authorities and their partners to take to embed participatory procurement practices
- Guidance, signposting and standards to support implementation
The first draft of these principles draws upon the research for ‘Stakeholder Engagement in Public Procurement for Artificial Intelligence: a Mission-Oriented Playbook’ and on an open online collaborative drafting process.
Over September, we’ve been testing the principles with governments, procurement experts and participation practitioners.
Session notes
The roundtable attracted a fantastic group of participants representing academics, consultants, procurement practitioners and civil society organisations, and over the course of an hour of discussions we identified a range of areas where participants said ‘yes, and’ or ‘no, but’ to the current draft.
Method statement: The session was recorded via Zoom and in-room recorder, and has been transcribed locally using MacWhisper, and then summarised using Google’s NotebookLM, with the summary reviewed by Tim against contemporaneous notes, and on the same day as the workshop while still fresh in mind.
Shared resources
In preparation for and during the session participants shared the following resources:
- Gear Shift - Driving Change in Public Sector Technology through Community Input, Meg Young et. al, Data & Society, 2025
- Key Considerations When Procuring AI in the Public Sector, Taraaz and Collaborative Research Center for Resilience, 2025
- Best practices in procuring AI: Lessons from Chile - Open Contracting Partnership, 2025
- Toolkit for Worker voice in public sector procurement of digital and AI systems in Wales, 2024
- Stakeholder Engagement in Public Procurement for Artificial Intelligence: a Mission-Oriented Playbook, Ismael Pena-Lopez 2025
- The Cyd ‘Procurement Reform Sandbox’: using foresight approaches to help deal with uncertainty and change, 2024
- Long Beach Co-Lab, 2025
- City of Seattle AI Plan, 2025
- From Trustworthy AI Principles to Public Procurement Practices, Merve Hickock, 2024
Feedback on the principles
• Preamble: The preamble establishes AI as infrastructure with long-term societal and environmental consequences, prioritising ethical, social, and democratic values over short-term gains, and aiming to shape AI use towards public value. While not directly commented on, the overarching theme of these principles being “strategic” was noted as a potential summary for all of them. The idea of principles being “anticipatory” to prevent issues was also highlighted, given the novelty of AI.
• Principle 1: Purpose (Defining a clear public mission)
One participant found this principle particularly interesting, questioning how participation can be brought into defining the purpose of a problem an AI procurement aims to address, who sets that problem, and the rationale for choosing an AI system. The point was also made that the purpose should ideally be agnostic about whether AI is the outcome, as current drives to procure AI can sometimes stem from government economic policies rather than local need.
• Principle 2: People (Including the right people from the start and throughout the contracting journey)
This principle resonated strongly with several participants, with one specifically drawn to it given their focus on participation for trustworthy AI. Others emphasised the importance of “people” because AI disproportionately affects certain communities, and mechanisms are needed to support and platform their voices in both AI development and procurement. It was underscored that people are at the heart of both solutions and challenges, necessitating collaboration to change mindsets, cultures, and behaviours. Some also noted potential overlap between this principle, “Partnership,” and “Inclusive Processes,” suggesting a need for refinement.
• Principle 3: Partnership (With the public, through consequence scanning, understanding who is affected, workers and publics)
One participant suggested a need to refine the distinction between this principle, “People,” and “Inclusive Processes” due to perceived overlap. The broader concept of workers’ voices in AI procurement was also linked to specific toolkits and union engagement examples.
• Principle 4: Participation (At all stages)
A participant critiqued this principle by stating it duplicates what is already done in standard procurement, arguing that the focus should be on aspects unique to AI procurement. The rapidly changing regulatory environment was identified as a critical, missing element that should be a separate item. It was also noted that while earlier discussions leaned towards pre-procurement engagement, the current work advocates for public involvement at all stages of the contracting journey.
• Principle 5: Informed Engagement (Build capacity for meaningful engagement, requiring access to vendor-neutral education and capacity building)
This principle was well-received, with a participant agreeing on the need for capacity building but proposing the addition of sub-principles for measuring capacity and readiness as essential for the implementation and enforcement of all 10 principles.
• Principle 6: Inclusive Processes (Embedding equity and diversity by design)
While acknowledging its importance, a participant questioned how to avoid replicating existing societal biases when AI learns from people, and how AI can be leveraged to course-correct for greater equity. The Long Beach CoLab was highlighted as a positive example of inclusive processes through compensated community engagement. Union engagement in Pennsylvania, which formally consults workers through a collaborative group, was also given as a concrete illustration. One participant noted that “inclusive processes” likely matters most in the current climate, and another suggested refinement due to overlap with “People” and “Partnership”. • Principle 7: Transparency and Oversight (Specifically on feedback and complaint mechanisms)
This principle garnered strong support, with one participant advocating for early transparency during the ideation and pre-procurement phases to reduce the burden on communities for providing input. The importance of feedback and complaint mechanisms was particularly emphasised as an often-overlooked aspect. This principle was also linked to the broader “transparency and accountability ecosystem”.
• Principle 8: Institutional Structures (For engagement, committing to not just ad hoc one-off processes)
A participant flagged this as a crucial principle, drawing on experience where institutional knowledge was lost when champions of initiatives moved on, underscoring the need for robust and lasting structures. The principle’s mention of “capacity for organizations” was also highlighted as relevant.
• Principle 9: Market Development (To diversify providers and partners)
This principle was viewed as important from a procurer’s perspective, but a question was raised about whether its focus should be more on “accessibility” than “development,” to mitigate potential conflicts of interest if government officials develop markets they later procure from.
• Principle 10: Evaluation and Learning (An ongoing process)
This principle was highly valued, with participants highlighting the importance of learning in a rapidly developing space.
Overarching Themes and Additional Considerations: Participants also raised several cross-cutting points, including:
• The fundamental challenge that many AI systems emerge without being formally procured, often through “backdoor implementations” or free offerings, necessitating a focus on the decision to use AI rather than just how it’s bought.
• The significant need for capacity building to equip government generalists to understand and manage AI.
• The importance of distinguishing what is truly “distinct” about AI that requires new principles from what can be integrated into existing regulatory tools, to avoid redundant “dual commitments”.
• The idea of applying these principles primarily to “high-impact” AI procurements, rather than all AI, to make them more practical and manageable.
• The recognition of AI’s connection to the broader context of digital public infrastructure (DPI).
• The need for multidisciplinary engagement beyond just procurement professionals, involving digital program managers, finance partners, and legal practitioners.
• The critical role of contract and relationship management throughout the AI lifecycle, especially to address risks like vendor lock-in from stealth implementations.
Session Hosts
The session was facilitated by Tim Davies from Connected by Data, and Warren Smith from Posterity Global on behalf of:
ParticipationAI explores the intersection of public participation and Artificial Intelligence (AI), focusing on integrating diverse civil society perspectives into the public procurement process. The ParticipationAI team, based across Spain, work rooted in feminist approaches, critical theory, and a politics of care.
Connected by Data is the campaign for communities to have a powerful voice in the governance of data and AI. Based in the UK, Connected by Data work to shift policies, practice and narratives to centre participatory, deliberative, collective and powerful governance of emerging technology.
The Open Government Partnership (OGP) is an organization of reformers inside and outside of government working to transform how government serves its citizens. OGP includes 74 countries and 150 local governments – representing more than two billion people – and thousands of civil society organizations.
The Open Contracting Partnership (OCP) working to transform procurement from an overlooked, underfunded paper-based chore to a digital, data-driven government service that can be an engine for innovation, sustainability and economic inclusion.