Coders Beyond Borders

A successful Global Dialogue on AI Governance should lead to concrete, actionable commitments that move beyond principles into implementation, especially for underserved communities. First, it should establish shared minimum standards for inclusive, human-centered AI that prioritize access, fairness, and usability for people who are often excluded from digital systems. Second, it should create pathways for local organizations and community-led initiatives to actively contribute to AI governance, not just as beneficiaries but as co-creators. Third, success would mean bridging the gap between policy and real-world impact. Many communities, including refugees and migrants, face barriers not because AI does not exist but because it is not accessible, understandable or designed with them in mind. Governance must address this gap. Finally, the Dialogue should result in mechanisms for sustained collaboration across governments, civil society, academia, and grassroots actors. This includes funding structures, knowledge-sharing platforms and open infrastructures that enable scalable, ethical AI solutions. From our perspective, success is when AI governance directly improves people's ability to understand their rights, access services and participate meaningfully in society.

• AI capacity-building
• Safe, secure and trustworthy AI
• Protection and promotion of human rights
• Open-source software, open data and open AI models

AI capacity-building is essential to ensure that communities, especially refugees and underserved groups, can meaningfully access and benefit from AI systems. Without this, AI risks reinforcing existing inequalities. Protection and promotion of human rights is central to our work. AI systems increasingly mediate access to legal information, services and opportunities. Ensuring that these systems uphold fundamental rights such as access to justice and non-discrimination is critical. Safe, secure and trustworthy AI is critical to ensure that systems provide accurate and reliable information, especially in high-stakes areas like legal guidance. Strong safeguards and validation are needed so users can trust AI without risk of harm or misinformation. Open-source software, open data, and open AI models enable collaboration, adaptability and localization. For community-driven initiatives like ours, open ecosystems allow us to tailor solutions to different linguistic, cultural and legal contexts while maintaining accessibility and scalability.

One critical issue not sufficiently captured is accessibility of AI in practice, not just availability. Many AI systems exist but they are not usable for people facing language barriers, low digital literacy or complex legal environments. Another key issue is community-led AI governance. Current discussions are often dominated by governments and large tech actors, while grassroots organizations working directly with affected populations are underrepresented. These actors bring essential insights into real-world challenges and should be structurally included in governance processes. Additionally, AI for access to justice remains underexplored. As legal systems become more complex, AI has the potential to democratize legal knowledge but without proper governance, it can also spread misinformation or create dependency on unreliable systems. Finally, cultural and linguistic inclusivity in AI systems is still insufficiently addressed. AI must go beyond translation to ensure contextual understanding and relevance across diverse communities.

In our sector, one of the biggest governance gaps is the lack of standards for accessible and reliable AI-driven legal information. Refugees and asylum seekers often rely on fragmented, unclear or outdated information. AI could solve this but without clear governance frameworks, there is a risk of misinformation, lack of accountability and inconsistent quality. Another challenge is the lack of inclusion in AI development and governance. Communities most affected by AI systems are rarely involved in their design, leading to solutions that do not reflect real needs. At the same time, there are strong opportunities. AI can significantly improve access to rights, services and opportunities when designed responsibly. For example, tools that simplify and translate legal information can empower individuals to participate more actively in society. There is also growing momentum for collaboration between civil society, academia and public institutions, which can lead to more grounded and impactful governance models.

The AI Dialogue can act as a bridge between global policy and local implementation. It should not only define principles but also facilitate collaboration mechanisms that allow these principles to be applied in diverse contexts. It can also serve as a platform to align standards across countries, especially in areas like human rights, transparency and accountability, while allowing flexibility for local adaptation. Importantly, it should elevate community-led and grassroots initiatives, ensuring that governance is informed by real-world challenges and lived experiences.

The Dialogue should build on existing collaborations between international organizations, academic institutions and civil society networks, such as refugee-led networks and partnerships with universities and legal institutions. These initiatives already demonstrate how AI can be applied responsibly in areas like access to justice, education, and inclusion. The added value of the AI Dialogue would be to connect these fragmented efforts, provide shared frameworks and enable scaling through funding, policy alignment and knowledge exchange.

Stakeholders should contribute not only through formal representation but also through structured, continuous engagement. We recommend a multi-layered format: High-level policy discussions Thematic working groups Community-led sessions and case studies Local organizations and affected communities should be supported to participate meaningfully, including through funding and capacity support.

Refugees, migrants and underserved communities are significantly underrepresented in AI governance discussions, despite being directly affected by AI systems in areas like legal access, employment and public services. Grassroots organizations and community leaders working with these groups are also often excluded. To include them: Provide funding for participation Create dedicated spaces for community input Recognize lived experience as expertise

Effective formats include: Co-creation labs where policymakers and communities design solutions together Scenario-based simulations to test governance approaches Hybrid dialogues combining online and local in-person sessions Open calls for community-led case studies and solutions These formats move beyond passive consultation toward active collaboration.

Effective AI governance should combine ethical principles with practical implementation. Examples include: AI systems that simplify and translate complex legal and administrative information into accessible formats Human-in-the-loop models where AI is combined with expert and peer support Open and modular AI infrastructures that allow localization and adaptation Community-based training programs that improve digital and AI literacy Approaches that integrate technology with human support systems are particularly effective in high-stakes contexts such as access to justice and social inclusion.