This research is a collaboration between Clare Dillon, a doctoral researcher at Lero, the Research Ireland Centre for Software, University of Galway, whose work examines how organizations enable, sustain, and govern InnerSource as organizational knowledge infrastructure, and Johan Linåker, senior researcher at RISE Research Institutes of Sweden, who has published extensively on open source in the public sector, including research on government OSPOs, open data ecosystems, and public sector open source collaboration. The framework builds on and extends this prior body of work on public sector OSPOs.
The study draws on interviews and structured engagement with government OSPO representatives across multiple jurisdictions and practitioners from the InnerSource Commons Foundation community. Factors examined include policy mandates, reuse demand from peer agencies, sovereignty and resilience considerations, security review capacity, procurement context, and the collaboration overhead of documentation, governance, and community management. The framework addresses a gap between macro-level policy advocacy for open source in government and the project-level realities of deciding what to open, what to InnerSource, and what to keep closed.
The session is designed to be interactive: after presenting the emerging framework, we will invite participants, particularly those from public sector and policy backgrounds, to test it against their own organizational contexts. This feedback will directly inform the next phase of the research, in which the framework will be validated with additional government OSPOs.

We envision this submission under Theme #2 (Governance). It examines how open models are positioned within state-led responses to concentrated AI infrastructural power, and it speaks directly to the OTR research agenda’s digital sovereignty questions—in particular, the conditions under which open technology delivers on the sovereignty promise without creating new forms of dependency, and the adequacy of existing arrangements as proxies for underlying public interest goals. The paper focuses on the AI stacks being assembled under South Korea’s sovereign foundation-model program, Brazil’s PBIA, India’s IndiaAI Mission, and the EU’s tech sovereign cloud procurement and funded model programs. Sources include national strategy documents, procurement and tender records, and the technical documentation of open-weight releases and their surrounding software. The analysis yields layer-specific criteria that stakeholders can apply to locate dependence in their own stacks and to assess where legal and contractual arrangements materially reduce dependence and where they merely relabel it.

1. **From Diagnosis to Institutional Design**
In alignment with the OTR Symposium’s theme “Shaping the Open Transition,” this session moves from documenting algorithmic risks to establishing protective structures. For civic tech advocates and public administrators, the primary barriers to open technology deployment are structural, procedural, and cultural rather than purely legal or technical. This presentation translates frontline field observations into clear, operational policy recommendations that decision-makers can readily integrate into public institutional strategies.

2. **Core Dilemmas Addressed**
**The Anti-Capture Problem:** Designing workflows to insulate public sector open AI ecosystems from vendor lock-in when underlying compute infrastructure and foundational models remain highly centralized.
**Civic Co-Stewardship Workflows:** Creating clear organizational pathways for civil society and open-source developer communities to hold formalized oversight roles within public sector deployments.
**The Strategic Resilience Metric:** Translating the structural, political, and digital sovereignty value of open civic infrastructure into frameworks that speak directly to budget units, procurement and governance.

3. **Empirical Context:** Establishing Baseline Challenges
To ground our policy recommendations, the session references structural operational lessons observed across critical public sector systems:
– **Resource Accountability (Hello Tractor Ecosystem — Nigeria/Kenya):** Examining the vulnerabilities that surface when proprietary tech aggregators enclose community-driven agricultural data pipelines.

– **Civic Health Infrastructure (DHIS2 Networks — Multi-National):** Evaluating how multi-stakeholder steering committees balance state oversight with distributed community development to protect public health data sovereignty.

– **Sovereign DPI vs. Enclosure (MOSIP Registries — Ethiopia/Togo/South Africa):** Analyzing the institutional risks of proprietary biometric vendor lock-in versus modular public digital infrastructure.

**4. The Public-Commons Partnership (PCP) Policy Roadmap**
The presentation shares three operational blueprints designed for public administrators:
**Data Reciprocity Guardrails:** Clear data-sharing criteria ensuring public sector training data remains open and protected from exclusive corporate monetization.
**Tripartite Institutional Architecture:** A clean governance model separating oversight into three pillars: State Ministries (legislative), Technical Stewards (code maintenance), and Civic Auditor Networks (civil society oversight).
**Resilience Procurement Criteria:** Valuation matrices helping budget units weigh long-term platform adaptability and independence against upfront software licensing fees.

**5. Presentation Structure**
Introduction: Why passive open licensing fails to protect public governance against AI monopolies.
Field Insights: Documenting friction points observed across African digital public infrastructure.
The PCP Roadmap: Implementing Tripartite Committees, reciprocity guardrails, and resilience procurement matrices.
Call to Action: Proposing a transnational Open Civic Tech Observatory to formalize North-South research pipelines.

Full author list (ORCID iD in parentheses):
Pen-Yuan Hsing (0000-0002-5394-879X), Julieta Arancio (0000-0002-8624-4182), Moritz Maxeiner (0000-0001-9946-3010), Robert Mies (0000-0003-0195-7405), Tim Landgraf (0000-0003-4951-5235), Julien Colomb (0000-0002-3127-5520)

This 20-minute session is structured in four parts, moving from problem to mechanism to evidence.

(1) Framing the gap (5 min). Two 2025 case studies anchor the talk: the Typhoon Danas asbestos-waste crisis, where a lone volunteer engineer built a tracking tool overnight but remained locked out of the command decisions his data could have shaped; and the Guangfu landslide dam overflow, where 500,000 volunteer visits and citizen-built platforms such as “Guangfu Superman” absorbed coordination failures that formal government information channels could not resolve.

(2) The institutional mechanism (7 min). We present Taiwan’s proposed “IT Squad,” a standing unit embedded inside Emergency Operations Centers through an MOU between central and local governments, pairing civic “Information Advisors” with government liaisons under clear activation thresholds and command boundaries. We then detail the Disaster Resilience Civic Tech Talent Pool: a vetted, publicly accountable registry offering per-diem pay, insurance, annual renewal, and removal for inactivity — the first attempt to price and formalize crisis-time civic contribution in Taiwan.

(3) The business-model innovation (5 min). The session’s core contribution is the Shared Component Library: government procurement funds the conversion of one volunteer-built disaster platform into twelve independent, MIT-licensed modules, each passing third-party SAST/SCA security review and shipping with a software bill of materials before release on public code repositories. This model reverses the usual sequence — instead of asking volunteers to sustain unfunded infrastructure indefinitely, government procurement retroactively funds the productization and maintenance of tools proven during a real emergency, while keeping the output as a public good under an open license.

(4) Evidence and open questions (3 min). We share preliminary findings from interviews with Taiwanese government, agencies and civic tech communities. We close with questions for the audience: how should contributions be priced fairly across volunteers with unequal risk exposure; what safeguards prevent procurement-funded open technology from being captured by contractors; and how transferable is this MOU-based, pre-authorized entry model to jurisdictions without Taiwan’s civic tech density.

The speaker leads the civic-participation work stream of Taiwan’s national digital resilience initiative and can bring primary interview data, draft policy instruments (MOU template, talent-pool selection criteria), and the component library’s public repository as supporting materials for the session.

This talk presents the ongoing work of the Research & Policy team at the Sovereign Tech Agency to develop an impact evaluation framework for public investment in open source software. We discuss the design and methodological choices underlying the framework (Goal-Question-Metric and Generalized Synthetic Controls) and the broader research and infrastructure challenges that surfaced while building it.

The target audience for this talk is researchers, policymakers and practitioners interested in measuring the impact of funding on OSS. Drawing on the experience of building this framework at the Sovereign Tech Agency, we highlight the broader challenges of building the research capacity needed for robust impact evaluation in this domain. Such methods extend beyond defining metrics: they rely on research infrastructure built on shared datasets, interoperable metadata, open tools, and collaborative research networks.

Impact evaluation in OSS is still an emerging field, with more open questions than answers. We will share what has worked and what has not in applying GQM and synthetic control methods to OSS funding data, and want to open a discussion with attendees. How else can responsible, community-aware evaluation be conducted? What other novel methods and data sources could be used to solve the aforementioned challenges?

Time outline of the presentation (20 min):

– 2 mins – Introduction: who we are, what the talk covers
– 3 mins – Policy context. Why is impact evaluation for OSS funding important?
– 5 mins – Sovereign Tech Agency’s approach
– 2 min – Qualitative approach and the need to complement it with data
– 3 min – Quantitative approach we are building, data and methods
– 3 mins – Challenges encountered developing these methods
– 2 mins – Conclusions and takeaways
– 5 mins – Q&A

This paper explores how digital participation platforms live, adapt, and sometimes fail within complex
urban and political ecosystems, synthesising comparative action research conducted through the development of Decidim-based participatory infrastructures across multiple European contexts, combining insights from platform governance, accessibility, intersectionality, democratic capabilities (Sen, 1999; Nussbaum, 2011), and participatory design. Particular attention is given to the digital translation of non-discursive democratic methodologies, including Legislative Theatre (Boal, 1992; Rubin, 2021), Forum Theatre, sociodrama (Moreno, 1953), photovoice (Wang & Burris, 1997), mapping, visual deliberation and other arts-based approaches that enable embodied, spatial and emotional forms of political expression often excluded from text-centred participation platforms.

The paper argues that the next generation of open participation technologies should not merely optimise digital consultation but support richer forms of democratic choreography, where online and offline participation, facilitation, accountability mechanisms and creative collective intelligence become part of an integrated participatory ecosystem. It concludes by proposing Political User Experience as a conceptual and practical framework for designing digital participatory policymaking systems that foster democratic capabilities, inclusion, institutional trust and meaningful behavioural adoption.

This research is joint work by Clare Dillon (Lero, University of Galway), Dr Lorraine Morgan (Lero, University of Galway), and Dr Daniel Izquierdo Cortazar (Bitergia). Clare Dillon is the presenting author. The study draws on qualitative interviews conducted across ten organizations spanning multiple sectors and geographies, including public sector, analyzed using Activity Theory to surface both the visible work of ISPOs and the systemic tensions that shape their effectiveness.

ISPO activities identified include creating resources and tooling, supporting practitioners, engaging leadership, and building community capacity. Common blockers include cultural resistance, compliance rules designed for proprietary development, and limited resourcing or executive sponsorship. Through the Activity Theory perspective, these tensions are interpreted as contradictions between organizational elements such as rules, tools, community, and division of labor, revealing why adoption stalls even where strategic commitment exists.

The empirical mapping of the ISPO/OSPO relationship is, to our knowledge, the first of its kind, and should be of direct interest to the OSPO research and practitioner community: it clarifies where the two functions can share resourcing and where InnerSource enablement requires distinct capabilities, since ISPOs must build community, incentives, and legitimacy entirely within organizational boundaries, without the external ecosystems and norms OSPOs draw upon.

The findings may hold particular promise for organizations in early stages of open source maturity, including in the public sector, a direction we identify for future research. The research team combines doctoral and senior academic expertise at Lero, the Research Ireland Centre for Software, with industry research expertise in open source community metrics and analytics. Clare Dillon is Vice President of the InnerSource Commons Foundation and previously served as its Executive Director.

The Reagent Collaboration Network (Reclone) is a global partnership building a model for a distributed infrastructure that increases access to molecular biology reagents. This model combines four elements i) the co-development of open source biotechnologies; ii) active community and infrastructure building to enable decentralised distribution, production and adoption; iii) the creation of open access training resources and protocols for capacity building and education; and iv) engagement with enterprise to scale solutions, while retaining open technology and affordable access as the core mission.

Developing open biotechnologies involves identifying, co-developing and distributing open resources to manufacture useful reagents. For biological materials, patents and material transfer agreements (MTAs) are the most significant legal tools that restrict downstream freedoms, so Reclone undertakes IP screening, de novo synthesises DNA and distributes it under an OpenMTA, a legal tool that meets the Open Definition. Open legal frameworks are necessary but insufficient to generate open technologies that are improved, maintained and sustained over time and we also promote excellent documentation, technical standards and proactive community coordination.

To build out global community and infrastructure, Reclone operates a hub-and-node model in which Regional Hubs in Latin America, Africa and Southeast Asia serve as anchor institutions that can distribute to local researchers, coordinate training, adapt protocols to regional needs and facilitate compliance with local regulations. We also aim to leverage the collective capacity of this decentralised network to crowdsource solutions but moving users of open technologies to become contributors and developers is a known challenge that is exacerbated by the real costs of the necessary R&D. We are particularly interested in the role of reagent manufacturing companies whose value proposition is typically based on local manufacturing rather than proprietary technologies, which has the potential to generate a productive collaboration space e.g. collaborative work with NexEnzymes in Chile is enabling new open protocols, kits and methods that will benefit both their commercial offerings and the community.

In this presentation, we will examine Reclone as a case study for both open biomaterials as a nascent area of open technology, and the development of open research infrastructure outside of digital technologies. We will present the broader technology, policy and legal factors that shape our work, the development of our model including quantitative and qualitative impact data, and an account of our future directions including challenges for governance and sustainability that could inform other initiatives.

This presentation makes the case that public procurement is entering a structural shift, driven on one end by the need to uphold public trust in public spending, and on the other by trust technologies and cybersecurity innovations that make verification independent, cheap, and machine-speed. New models and affordances are possible as a result, and this presentation walks through the procurement capabilities coming into reach.

Many of the problems governments face today cannot be solved by any single actor, whether vendor, government, or otherwise, and require an “all hands on deck” approach to public problem-solving. Meeting them requires more open source ways of working in the public sector and the ability to coordinate across organizational boundaries (between departments, vendors, governments, and civic actors) at a scale traditional accountability mechanisms cannot support. Independent verifiability is the missing substrate as it lets parties who do not share an institutional trust relationship still work, build, and transact together efficiently.

With trust in national governments persistently low, more people across OECD countries now distrust their national government than trust it (OECD Trust Survey, 2026) there is mounting pressure to tie public spending to demonstrated results. Outcomes-based contracting answers this pressure but costs more to administer due to heavy compliance and reporting needs. Cheap, independent verification addresses this increased cost and can open the door for smaller vendors to enter and stay in public markets (previously fixed compliance costs that don’t scale with contract size have structurally excluded small but highly innovative suppliers). Policy is already moving in this direction with the US GSA’s FedRAMP 20x initiative shifting toward machine-readable compliance evidence, and the EU’s Single Procurement Document’s “once-only principle” for reusable supplier qualification, however these efforts lack a trust substrate strong enough to work across trust domains and be deployed more broadly.

Government-to-government (G2G) markets are new concept and emerging model. In-house building for digital infrastructure often outperforms traditional procurement on cost and effectiveness, and governments share remarkably similar needs, yet procure the same solutions over and over. In a G2G model, one administration builds in-house, then sells to and adapts for others to ease acquisition burden, align incentives, and build lasting capacity inside the public sector. Early examples include India Stack, now being explored as a foundation for Kenya’s digital infrastructure, and COVID-era technologies shared between municipalities.