The Wikimedia Foundation

A successful first Global Dialogue should develop implementation pathways that ensure that the development and deployment of AI benefits the communities contributing and governing the open data, open knowledge, and open code that AI systems depend upon. This includes reaffirming the Global Digital Compact's commitment to digital public goods (DPGs)—including open knowledge, open data, open-source software, and open AI models—as foundational building blocks for a more inclusive AI ecosystem that is multilingual and human-rights respecting. The Dialogue should also address emerging structural imbalances in the AI ecosystem. As AI systems increasingly rely on openly available, human-created knowledge, there is growing asymmetry between extraction and contribution, with large-scale reuse placing strain on public-interest digital infrastructure while reducing visibility and participation. Sustainable AI governance must therefore include norms for responsible reuse, attribution, and support for the digital commons.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models

The priorities selected focus on conditions required for people around the world to access and contribute to the digital commons - publicly accessible and freely licensed data, content, and code. The social, economic, cultural, linguistic, and technical implications of AI are central because free knowledge projects like Wikipedia and Wikidata underpin a global knowledge commons. AI systems increasingly shape how knowledge is produced, accessed, and represented, with significant risks of reinforcing existing inequities-particularly for underrepresented languages. Addressing these implications is essential to ensuring that AI supports knowledge equity and everyone's ability to participate in the creation of shared knowledge. Likewise, the protection and promotion of human rights such as freedom of expression, access to information, privacy, and participation in public life are all prerequisites for a healthy AI ecosystem. Community-led platforms like Wikipedia demonstrate that rights-respecting governance models are not only viable, but also effective at scale. Transparency, accountability, and human oversight are critical to maintaining trustworthy and verifiable information ecosystems, as well as supporting individuals in exercising the basic skills of civic engagement. The ability to monitor, debate, and revise the knowledge, as well as the community-developed policies that govern these shared online spaces, is essential to safeguard information integrity and ensure that AI systems remain adequately representative of and aligned with societal values. Open models empower independent researchers and third parties to examine systems for risks such as bias, vulnerabilities, or harmful outcomes. In addition, openness helps create a clear "paper trail" of how systems are built and perform over time, which makes it easier to reproduce results, scrutinize decisions, and continuously improve systems. Finally, open-source software, open data, and open AI models are core components required to cultivate a robust digital commons that is essential to the success of public interest projects like Wikipedia, civic tech projects like Decidim or OpenStreetMap, as well as driving the development of new innovations and digital public goods. Only when content can be legally shared, edited, and reused can seamless online collaboration and creation flourish and drive human rights and development. That openness must be paired with responsible governance by the affected communities.

A key issue is the sustainability of the open knowledge ecosystem in the AI era. AI systems rely on large-scale extraction of openly available data, often without attribution or contribution back to the ecosystems that produce and maintain this open knowledge. On Wikipedia, we see a significant rise in bots scraping our content at alarming scale. This activity is driven by commercial use, not individual use. Mass-scale scraping places undue strain on our technical infrastructure, including increases in bandwidth demand, pressure on core data centers, and grows our carbon footprint. The damage is twofold. It degrades the experience for everyday users, and it raises operational costs for maintaining projects that are meant to help humans, not bots, and serve the public interest, not commercial aims. This extractive dynamic also weakens participation incentives for the volunteer community that contributes and governs open knowledge projects. Addressing this requires a shift from viewing openness as unlimited access toward promoting fair and responsible reuse. This includes stronger norms around attribution, clearer referral pathways, and greater support for the underlying infrastructure that enables access. More sustainable access models already exist, including structured services that support high-volume use while reducing strain on core systems, such as Wikipedia's Enterprise API. Focusing on smaller, context-specific AI is an additional means to prioritize efficiency, relevance, and adaptability of AI systems and lessening the immense strain of large-scale AI development. Finally, safeguarding information integrity is increasingly urgent. AI systems both depend on and influence the global information ecosystem, making it essential to protect high-quality, human-curated knowledge sources and ensure transparency in how information is generated and presented.

CHALLENGES Limited transparency around how AI models work—especially which sources they use and what criteria shape outputs—reduces auditability and local adaptation. Recycling unverified AI-generated content into open knowledge ecosystems can accelerate misinformation, while misattribution or poor source representation undermines trust in credible public interest media. Many AI tools also fail across diverse linguistic and cultural contexts, reinforcing knowledge gaps and systemic biases. Governance gaps create major imbalances in who shapes AI's future. Smaller actors, including startups, nonprofits, and community-led projects, struggle to participate on equal footing given the speed and scale of AI development driven by large players. They often lack personnel, time, and resources required to engage in shaping standards, infrastructure, and market norms. Advanced AI systems frequently rely on openly available, human-created content and data. This places pressure on public-interest platforms, where automated extraction increases bandwidth demand, consumes resources, and diverts capacity from human contributors. These dynamics raise questions around equitable value sharing across the ecosystem. Left unresolved, community-led initiatives like Wikipedia face structural disadvantages in scaling and long-term sustainability despite their foundational role. OPPORTUNITIES Addressing these governance gaps presents an opportunity to strengthen a more inclusive and sustainable digital commons that provides the building blocks for more open and transparent AI tools that serve public interest objectives. For Wikimedia, this amounts to embracing open knowledge infrastructure as a shared public good: ensuring that high-quality, multilingual, human-curated knowledge remains accessible and sustainable as a foundation for AI systems. It also means supporting not only open-source code and content, but also technical infrastructure, contributor communities, and governance foundations needed to sustain these systems over time. Key actions include: strengthening interoperability through open standards; supporting shared data structures enabling responsible; privacy-preserving access; and investing in human infrastructure of maintainers, contributors, and community governance. The success of Wikipedia and Wikidata shows that collaborative approaches grounded in DPGs can lower participation barriers, support locally relevant innovation, and enable ecosystems where public institutions, civil society, and technical communities build shared infrastructure.

The AI Dialogue can play an important role by moving international cooperation from high-level principles to practical, shared approaches for building and governing AI systems. To do so, it should focus on enabling collaboration around common infrastructure and resources, particularly open data, open knowledge, open-source software, open AI models, and the contributor communities that sustain these assets. These provide a concrete foundation for cooperation by allowing countries and communities to adapt, audit, and deploy AI systems in ways that reflect their local contexts, while reducing dependencies on a small number of providers. A key contribution of the Dialogue would be to advance open knowledge as essential digital infrastructure that needs to be supported and sustained. High-quality, openly accessible knowledge systems underpin both AI development and information integrity, yet remain under-recognized and under-supported in global governance frameworks. Strengthening these systems is essential to closing the AI divide and ensuring more equitable participation in AI development. To meaningfully advance cooperation, inclusive and participatory formats must be prioritized, including dedicated spaces for community-led and public-interest perspectives. A focused discussion on the role of open knowledge infrastructure in closing the AI divide would be a valuable step toward operationalizing these goals.

The AI Dialogue should build on existing initiatives that provide practical frameworks for open, inclusive, and accountable AI development. The experience of community-governed knowledge systems such as Wikipedia and Wikidata demonstrate how open, participatory models can sustain high-quality, multilingual knowledge infrastructure at global scale, which is an increasingly critical foundation for AI. Both of these topics are directly reinforced by the Global Digital Compact, which recognizes digital public goods as essential to closing digital divides and advancing inclusive digital transformation. The Dialogue can add value by connecting these efforts and moving toward coordinated global action, including integrating DPGs into national AI strategies, promoting interoperability, and supporting investment in shared infrastructure and standards required to support open source communities. Building on lessons from the open source movement, it should also advance stronger public-interest investment in foundational AI resources—such as research, safety capabilities, and trusted data infrastructure—particularly where market incentives fall short. Strengthening these shared building blocks can help create a more balanced ecosystem in which countries and communities are able to develop, adapt, and govern AI systems in line with their own needs, ensuring AI governance is implementable, inclusive, and sustainable.

Make a call for proposals to include community-led sessions during the Global Dialogue. From the open knowledge movement's side, we would be pleased to co-chair a session with a member state on the role of the digital commons that support open-source AI in bridging the AI divide as part of the Global Dialogue in July.

Global Majority voices broadly speaking, open source developers and developers of open knowledge.

N/A

The Wikimedia Foundation's projects offer concrete, field-tested approaches to AI governance grounded in openness, human rights, and public-interest infrastructure. Wikipedia's community-led governance model demonstrates how transparency, human oversight, and participatory decision-making can operate effectively at global scale, while its verifiability and reliable sourcing standards help sustain a high-quality information ecosystem that is critical for AI systems relying on large-scale data. Wikimedia's use of open licensing, combined with structured access pathways such as APIs and services like Wikimedia Enterprise, illustrates how openness can be paired with responsible and sustainable reuse models, including attribution and contribution back. More broadly, Wikimedia projects function as knowledge infrastructure and digital public goods, underscoring the need to support shared resources that underpin trustworthy and inclusive AI. These approaches align with and complement broader multi-stakeholder efforts including participation at convenings like the Internet Governance Forum and our membership in the Freedom Online Coalition and the Digital Public Goods Alliance, both of which promote human rights-based, transparent, and accountable digital governance frameworks.