Rayznews

AI governance is increasingly shaped by discussions around values, but there remains a fundamental gap. The meaning, definition, and practical application of "AI values" vary widely across regions, institutions, and communities. Without a shared baseline, these differences risk deepening fragmentation rather than fostering trust. There is a clear need to move toward greater standardization, not as a rigid framework, but as a collaborative process built on mutual learning and contextual understanding. A genuine multi-stakeholder approach is essential in this process. Governments, private sector actors, technical communities, civil society, and academia must not only participate but actively listen to one another. More importantly, this approach must go beyond representation in name and ensure meaningful inclusion of voices that are often overlooked. Learning must be reciprocal, not hierarchical. At the same time, it is important to acknowledge the underlying politics that often operate under the banner of "community," "people," or "regional interests." These narratives can unintentionally marginalize those without access, influence, or platforms. Addressing this requires deliberate efforts to create space, amplify unheard perspectives, and recognize contributions from those working at the margins. From the perspective of underdeveloped countries, the challenge is not only about access to AI technologies but also about representation in shaping the values that govern them. There is a pressing need to support and elevate individuals and communities who are building pathways toward opportunity, innovation, and leadership in their local contexts. Creating inclusive AI systems is not just a technical goal. It is a shared responsibility to ensure that the future of AI reflects diverse realities, supports equitable growth, and empowers those who have long remained unheard. Shreedeep Rayamajhi, Nepal

• AI capacity-building
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models

My choices come from a simple concern: if AI is shaping the future, then more people should have the ability to shape AI. AI capacity-building matters because many communities, especially in developing countries, are still on the sidelines. Without the right skills, knowledge, and support systems, they don't get to participate meaningfully. For me, capacity-building is about giving people the tools and confidence to create, not just consume. It's about enabling local innovation and leadership that reflects real needs on the ground. Transparency, accountability, and human oversight are about trust. People need to understand how AI systems make decisions, who is responsible when things go wrong, and where humans step in. Without this, AI risks becoming distant and unaccountable. Keeping humans in the loop ensures that technology stays aligned with ethical values and real-world consequences. Open-source, open data, and open AI models are what make all of this more inclusive. Openness allows people from different backgrounds to learn, experiment, and contribute. It reduces dependency on a few powerful actors and creates space for collaboration and shared progress. Together, these priorities are about fairness. They are about making sure AI is not built by a few for a few, but shaped by many, including voices that are often unheard.

Yes, a few important issues still sit in the gaps. One of the biggest is who actually gets a seat at the table. We often talk about inclusion, but in reality, many voices especially from developing countries are still missing or only lightly represented. Being present is not the same as being heard. There needs to be space where perspectives from local communities genuinely shape decisions, not just appear in discussions. Another issue is the growing concentration of power in AI. A small number of actors control most of the data, infrastructure, and advanced systems. This affects everything from access to innovation to who benefits economically. For many countries, it creates dependency rather than opportunity, which is a concern that goes beyond technology into fairness and global balance. There is also the question of context. AI systems are often built far away from the places where they are used. As a result, they may not reflect local languages, cultures, or realities. This can make them less effective, or even harmful in some cases. Making AI work for people means grounding it in their everyday context. Finally, we cannot ignore the long-term impact. AI is already influencing jobs, education, and social systems, and it also has environmental costs. These are not separate issues, they are part of the same conversation about what kind of future we are building. At the core, all of this comes back to balance. AI should not just move fast, it should move fairly, with space for more voices, especially those that have been left out for too long.

In Nepal and much of South Asia, these governance gaps are not abstract, they show up in real ways. A major challenge is limited capacity. There is growing interest in AI, but many institutions simply don't have the skills, resources, or policies to guide its use. As a result, technologies are often adopted from outside without fully understanding how they work or how they should be managed locally. That creates risks, especially when there is little room to question or adapt them. Transparency and accountability are also still weak. There are no clear systems to check how AI is being used or to respond when something goes wrong. This becomes more concerning as AI starts influencing information, finance, or public services. People are expected to trust systems they don't fully understand. Another challenge is lack of representation. Countries like Nepal are rarely part of the spaces where global AI rules and standards are shaped. This means our realities, languages, and priorities are often missing from the bigger picture. At the same time, there is real opportunity. AI could help bridge long-standing gaps in areas like education, healthcare, agriculture, and disaster response. With the right support, local innovators can build solutions that actually fit our context. The rise of open-source tools and collaborative models also makes it easier to learn, experiment, and contribute, even with limited resources. So it's a mixed space. There are clear risks of being left behind or becoming dependent, but there is also a chance to build something more inclusive if the right steps are taken now.

The AI Dialogue can be a space where things actually start to connect. Right now, conversations on AI are often happening in separate rooms. Governments, tech companies, researchers, and civil society all have their own discussions, but they don't always come together in a meaningful way. The Dialogue can help bridge that gap by creating a space where people not only speak, but also listen and try to understand each other's realities. It can also make global discussions more inclusive. Many countries, especially from the Global South, are still missing from key decision-making spaces. The Dialogue has the potential to bring those voices in, not just as participants, but as contributors whose experiences shape the direction of AI governance. Another important role is learning from each other. Instead of staying at the level of big ideas, the Dialogue can focus on what is actually happening on the ground. Sharing real experiences, both successes and failures, can help others adapt and move forward more confidently. It also helps build trust. When there is open and continuous conversation, it becomes easier to understand intentions and reduce misunderstandings. This matters in a space where there is both cooperation and competition. At the end of the day, the AI Dialogue can help bring some balance. It can connect different perspectives, reduce gaps, and move us a little closer to a shared approach where more people feel included in shaping the future of AI.

The AI Dialogue doesn't have to reinvent the wheel. There are already many initiatives working on AI governance that it can build on. Globally, organizations like the United Nations, OECD, and UNESCO are setting principles and ethical guidelines for AI. At the same time, technical and community-driven groups—like ICANN, the Internet Society, and regional forums such as the Asia-Pacific Regional Internet Governance Forum—show how multistakeholder approaches can work in practice. Open-source and collaborative AI projects are also creating ways for people everywhere to contribute and learn. The problem is that these efforts often operate in separate spaces. Policy discussions, technical communities, and civil society rarely fully connect, which means lessons and experiences don't always travel. This is where the AI Dialogue can make a difference. It can act as a bridge, linking these initiatives so that knowledge and good practices are shared, not siloed. It can also make discussions more inclusive, giving underrepresented regions and voices a consistent platform to shape decisions rather than being occasional participants. Importantly, the Dialogue can focus on practical action. Instead of creating new rules, it can help translate existing principles into steps that work on the ground, encourage collaboration between different groups, and help everyone align on shared goals. In short, the AI Dialogue can bring what already exists together, amplify voices that are often missing, and make global AI governance more connected, fair, and grounded in reality.

Every stakeholder has something different to bring to the table, and the AI Dialogue works best when all of those voices are included. Governments can share how they're thinking about AI rules, policies, and public interest priorities. Private sector actors bring hands-on experience what works, what doesn't, and the real challenges of building and scaling AI systems. Civil society and community groups represent the people who are often left out, pointing out social impacts, ethical concerns, and local needs. Researchers and academics provide evidence, data, and lessons from studies that help ground the discussion in reality. To make the Dialogue effective, the format should encourage active participation and practical outcomes. # Panels with diverse voices: Make sure each discussion includes policymakers, tech experts, community representatives, and researchers so the conversation is balanced. # Interactive workshops: Real-world examples and hands-on problem-solving help participants learn from each other, not just talk about ideas. # Regional sessions: Small breakout groups focused on local issues can feed into the larger global discussion. # Focused working groups: Groups on ethics, open AI, or capacity-building can create concrete recommendations. # Ongoing engagement: An online platform or community space lets participants continue the conversation, share resources, and track progress between events. The goal is simple: create a space where everyone listens as much as they speak, where experiences are shared, and where practical, inclusive outcomes emerge. It's about turning conversation into action and ensuring that the future of AI reflects the needs and realities of all people, not just a few.

Many important voices are still missing in global AI governance conversations, and that matters because AI affects everyone. First, countries from the Global South, like Nepal and much of South Asia, are often on the sidelines. They face unique challenges with AI but rarely get to shape international rules and standards. Second, local communities and marginalized groups women, rural populations, and economically disadvantaged people are frequently absent from discussions, even though they are the ones most affected by AI in healthcare, finance, or public services. Third, smaller technical communities and grassroots innovators often have practical know-how but little access to global forums. Their insights could make AI systems more grounded and effective. To include these voices, we need more than symbolic participation. Some ways to do this could be: # Support participation with funding, translation, and preparatory resources so people from underrepresented regions and communities can join discussions meaningfully. # Regional and local dialogues that feed into the global conversation, bringing local challenges and lessons to the international stage. # Mentorship and capacity-building to help smaller civil society groups and local innovators confidently engage in technical and policy debates. # Digital platforms where voices can contribute continuously, not just at formal events. # Inclusive decision-making mechanisms to ensure these contributions actually shape outcomes, not just sit on the sidelines. Including these perspectives isn't just fair—it makes AI governance stronger. When people closest to the impacts of AI have a say, systems become more equitable, practical, and relevant to real lives.

To make the AI Dialogue truly engaging, it needs formats that encourage conversation, collaboration, and real learning not just people sitting and listening. Interactive panels and roundtables work well when participants from different backgroundspolicymakers, tech experts, civil society, and local innovators can share their perspectives and respond to questions in real time. It turns a discussion into a conversation. Hands-on workshops are also powerful. Small groups can explore real-world scenarios, like AI in healthcare or education, identify challenges, and brainstorm solutions together. It's a practical way to learn from each other. Regional breakout sessions give space for local contexts to be heard. Participants can discuss challenges and opportunities specific to their region, and those insights can feed into the bigger global conversation. Thematic working groups can dig deeper into focused topics like ethics, transparency, capacity-building, or open AI, producing recommendations that the larger group can adopt. Digital and hybrid tools help keep everyone involved, even if they can't attend in person. Online forums, collaborative platforms, and live polls allow people to contribute ideas and keep the conversation going beyond the event. Storytelling sessions that highlight lived experiences make the discussion real. Hearing from communities affected by AI reminds everyone of the human side of these technologies. Finally, mentorship and peer-learning circles let less experienced participants learn from experts while sharing their own insights, creating a two-way exchange instead of a top-down flow. Using these approaches together can make the AI Dialogue more inclusive, dynamic, and action-oriented. The goal is for participants to not just talk, but to listen, learn from each other, and co-create solutions that really work in the real world.

https://www.researchgate.net/publication/400802151_Digital_Public_Infrastructure_in_Nepal_Progress_Problems_and_the_Path_to_2030 https://www.researchgate.net/publication/383813258_Global_Digital_Compact_GDC_Its_Impact_on_Nepal https://www.researchgate.net/publication/384540646_Recommendation_for_the_Concept_Paper_on_the_use_and_practice_of_Artificial_Intelligence_AI_By_Shreedeep_Rayamajhi https://www.researchgate.net/publication/371349829_AI_Governance_Framework_for_Nepal_by_Shreedeep_Rayamajhi