SpaceAI

Honestly, I think success for the first Global Dialogue on AI Governance is not about producing a perfect framework. It is about setting the right direction early, before things become too fixed. For me, a successful outcome would be one where the conversation moves beyond the usual voices. Right now, most AI systems are shaped by a few regions, yet they are used globally. If this dialogue can genuinely bring in perspectives from places that are building under constraints, then it is already doing something important. Another outcome that would matter is recognizing that data is not just a technical input, it is infrastructure. Many countries are not lacking ideas or talent, they are lacking access to quality, representative data. If the dialogue can push for global commitment around data access, data sharing, and inclusion of underrepresented languages, that would be a big step. It would also be important for the dialogue to stay practical. Not just principles, but direction that countries and builders can actually use. Things like how to balance innovation and regulation, how to support local ecosystems, and how to avoid widening the gap between those who build AI and those who only consume it. Lastly, success would mean creating continuity. This should not be a one-time conversation, but the start of an ongoing process where different regions continue shaping how AI evolves. If this dialogue can shift who is included, what is prioritized, and how seriously these issues are taken, then it would be a strong foundation.

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

From my perspective, these priorities reflect both what I am currently working on and the gaps I am seeing while building AI solutions in a low-resource environment. AI capacity-building is critical because the challenge is not just access to AI tools, but the ability for people and institutions to actually understand, build, and adapt these systems to their own contexts. Without that, many regions will remain users of AI rather than contributors. The broader social, economic, cultural, and especially linguistic implications of AI are also very important to me. Most AI systems today do not fully reflect the diversity of languages and realities across the world. This creates a risk where entire communities are left out of the benefits of AI or misrepresented within it. Open-source software, open data, and open AI models are important because access remains one of the biggest barriers. In many cases, the issue is not lack of innovation, but lack of access to the resources needed to build. Open ecosystems create more opportunities for participation and local innovation. Finally, transparency, accountability, and human oversight are essential to building trust. As AI systems become more embedded in everyday life, people need to understand how decisions are made and have confidence that these systems are fair and responsible. Together, these areas reflect a need to make AI more inclusive, accessible, and grounded in real-world contexts.

Yes, I think one important cross-cutting issue that is not fully captured is the role of data as infrastructure. In many discussions, data is treated as a technical component of AI systems, but in reality it functions more like infrastructure. Without access to high-quality, representative, and locally relevant data, it is very difficult for many regions to meaningfully participate in AI development. This creates a situation where some countries build and shape AI systems, while others mainly consume them. Closely linked to this is the issue of data inequality. Large datasets are concentrated in a few regions and in a few languages, which means that many communities are either underrepresented or not represented at all. This affects not only performance, but also fairness and relevance. Another emerging issue is language inclusion. AI systems are advancing quickly, but support for many local and indigenous languages is still limited. This creates a risk of digital exclusion, where people are unable to fully interact with or benefit from AI technologies in their own languages. There is also a need to better recognize contextual deployment challenges. Many governance discussions assume a level of infrastructure, connectivity, and institutional capacity that does not exist everywhere. This creates a gap between global frameworks and local realities. These issues are interconnected. Addressing them would help ensure that AI development is more inclusive and that more regions can actively contribute to shaping its future, rather than being left out of it.

From my perspective, the governance gaps in these areas are already shaping how AI is developing in my country and across the region. One of the biggest challenges is around AI capacity-building. There is growing interest in AI, but limited structured support in terms of skills development, infrastructure, and institutional readiness. This slows down the ability of local innovators to build and deploy solutions at scale, even when there is clear demand. Another major issue is around data and openness. There is very little access to high-quality, locally relevant datasets, and most existing AI models are not trained on our contexts or languages. This creates a gap where solutions either do not perform well locally or have to be heavily adapted, which increases cost and complexity. At the same time, it highlights an opportunity to build local data ecosystems and contribute new datasets that reflect our realities. The linguistic and cultural dimension is also significant. Many communities are not represented in current AI systems, especially in terms of language. This limits access and reduces trust in these technologies, but it also creates an opportunity to develop more inclusive systems that better reflect local needs. On the side of transparency and accountability, there is still limited awareness and few clear guidelines on how AI systems should be evaluated or governed locally. This can slow adoption, but also gives space to build responsible frameworks from the ground up.

The AI Dialogue can play a very important role in creating alignment across regions that are currently moving at very different speeds in AI development. Right now, a lot of AI governance conversations are happening in silos, often driven by a few countries or regions. The Dialogue creates an opportunity to bring those perspectives together and make sure that global approaches are not shaped by only a small part of the world. It can also help surface practical realities from different contexts. For example, what works in highly resourced environments may not translate well in regions with limited infrastructure or data. By bringing these differences into the conversation early, the Dialogue can support more balanced and adaptable governance approaches. Another important role is in setting shared priorities. Even if countries take different regulatory paths, there should be some common understanding around inclusion, access, and responsible development. The Dialogue can help define those shared principles while still allowing flexibility. It can also strengthen collaboration beyond governments, especially with builders, researchers, and local ecosystems who are directly working with AI systems. Including these voices makes the outcomes more grounded and actionable.

There are already several initiatives working on different aspects of AI governance, and the Dialogue can benefit from connecting and aligning with them rather than starting from scratch. For example, frameworks developed around AI ethics and responsible AI, as well as ongoing work in open-source and open data communities, provide a strong foundation. There are also regional and international collaborations focused on AI capacity-building, research, and policy development that the Dialogue can learn from. At the same time, many of these efforts are fragmented. They often operate in parallel without enough coordination, and some regions are still underrepresented in these spaces. The added value of the AI Dialogue is its ability to act as a unifying platform. It can bring together these different efforts, identify what is working, and help scale those approaches across regions. It can also highlight gaps, especially in areas like data access, language inclusion, and local capacity, which are not always fully addressed in existing initiatives. Another important contribution would be amplifying voices that are not always included in global conversations, particularly from low-resource environments. This would help ensure that future governance approaches are more inclusive and better reflect global realities.

Different stakeholders need to be engaged in ways that reflect how they actually work, not just through formal submissions. Governments can contribute through policy direction and national priorities, but it is equally important to involve people who are building and deploying AI systems. Builders, researchers, and startups should be able to share practical insights based on real implementation challenges. Civil society and community organizations also play an important role, especially in highlighting how AI affects everyday life, access, and inclusion. Their input helps ground the conversation beyond technical and policy perspectives. In terms of structure, the Dialogue should combine written submissions with more interactive formats such as regional discussions, small working groups, and sector-specific sessions. This would allow deeper engagement rather than only high-level contributions. It would also help to create feedback loops, where contributors can see how their input is being used or reflected. This builds trust and encourages continued participation.

Many of the voices currently underrepresented are those from low-resource and emerging regions, where AI is being adopted under very different conditions. This includes local builders, small startups, and researchers who are working with limited data, infrastructure, and funding. Their experiences are often missing, even though they reflect the realities of a large part of the world. Language communities are also underrepresented. Many global discussions happen in a few dominant languages, which limits participation and excludes perspectives tied to local contexts. To improve inclusion, there needs to be more intentional outreach and support. This could include regional consultations, partnerships with local ecosystems, and making participation accessible in multiple languages. It is also important to lower barriers to entry. Not everyone can engage through formal policy language, so there should be flexible ways for people to contribute their insights.

I think engagement needs to move beyond traditional formats and create space for more practical and context-driven contributions. One approach would be to organize regional and thematic workshops where participants can discuss specific challenges and share real use cases. This allows for deeper conversations compared to general submissions. Another format could be case-based contributions, where participants describe real-world examples of how AI is being built or used in their context. This would bring more practical insight into the Dialogue. Interactive sessions such as roundtables or small group discussions can also help create more dynamic engagement, especially when they bring together different types of stakeholders. It would also be useful to include digital platforms that allow ongoing contributions, feedback, and discussion over time, rather than limiting participation to a single deadline. Also, incorporating multilingual engagement formats would make participation more inclusive and reflect a wider range of perspectives.

In my opinion the most effective examples of AI governance are the ones that combine clear principles with practical implementation, especially around data, transparency, and inclusion. One strong example is the growing use of open-source AI ecosystems, where models and tools are made more accessible to researchers and developers. This helps reduce concentration of power and allows more regions and smaller teams to participate in AI development, even if they do not have large infrastructure. Another important approach is the development of AI ethics and accountability frameworks within organizations. These frameworks focus on transparency, human oversight, and bias mitigation. In practice, they help ensure that AI systems are reviewed before deployment and continuously monitored for unintended impacts. There are also emerging data governance policies that focus on privacy, data protection, and responsible sharing of data across borders. These are important because data is the foundation of AI systems, and without clear rules, it can easily lead to misuse or exclusion. In some regions, AI capacity-building programs and innovation hubs are also making a difference. These initiatives support local developers, researchers, and startups to build context-relevant solutions and improve technical skills. This helps close the gap between global AI development and local application. Further, we could also use platforms that encourage multi-stakeholder collaboration, where governments, private sector, and civil society engage together, are particularly effective. They help ensure that governance is not developed in isolation, but reflects different needs and perspectives.