Ministry of Education

A successful Global Dialogue on AI Governance should result in practical, inclusive, and actionable frameworks that move beyond discussion into implementation. First, the dialogue should establish a shared global baseline for responsible AI, including principles for safety, transparency, accountability, and human oversight that are adaptable across different national contexts. This baseline must be usable not only by governments, but also by educators, developers, and small innovators. Second, success would include concrete capacity-building commitments, particularly for developing nations. Many regions, including smaller economies, risk being left behind in AI adoption. The dialogue should therefore produce initiatives that support equitable access to AI tools, infrastructure, and training. Third, it should foster multi-stakeholder collaboration, connecting policymakers, educators, researchers, and industry practitioners. Real-world AI systems—especially in education—require alignment between policy and practice, and this dialogue should create channels for continued collaboration beyond the event. Fourth, meaningful success would involve guidelines for ethical and human-centered AI deployment, particularly in sensitive sectors such as education. AI systems must enhance human potential, not replace or marginalize it. Finally, the dialogue should produce a clear roadmap with measurable milestones, ensuring that discussions translate into sustained global action rather than remaining aspirational. In essence, success lies in bridging the gap between global principles and local implementation, ensuring that AI benefits are accessible, responsible, and inclusive for all.

• Safe, secure and trustworthy AI
• AI capacity-building
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

These priorities reflect the need to balance innovation with responsibility, particularly in real-world applications such as education. Safe, secure and trustworthy AI is fundamental. Without trust, adoption will be limited, especially in public sectors like education where systems directly impact learners. Ensuring robustness, reliability, and protection from misuse is essential. AI capacity-building is equally critical. There is a growing divide between countries and communities that can effectively leverage AI and those that cannot. As an educator, I see firsthand how access to AI tools and training can transform learning outcomes. Building capacity ensures that AI benefits are distributed more equitably. Protection and promotion of human rights must remain central. AI systems should respect privacy, prevent bias, and uphold dignity. In education, this includes safeguarding student data and ensuring fair, inclusive learning experiences. Transparency, accountability, and human oversight are necessary to maintain control and responsibility over AI systems. Users must understand how decisions are made, and there must be clear accountability mechanisms when systems fail or produce unintended outcomes. Together, these priorities support a human-centered AI ecosystem that is both innovative and ethically grounded, ensuring that technological progress aligns with societal values.

One important emerging issue is the rise of agentic AI systems-autonomous, multi-agent systems capable of planning, decision-making, and continuous learning. These systems introduce new governance challenges, particularly around accountability, control, and unintended behavior. Existing frameworks may not fully address the complexity of such systems. Another critical issue is the integration of AI into education at scale. While AI has the potential to personalize learning and improve accessibility, there is a risk of over-reliance, reduced critical thinking, and widening inequalities if implementation is not carefully guided. Additionally, the digital divide in AI readiness remains a major concern. Beyond access to tools, disparities exist in infrastructure, data availability, and local expertise. Governance discussions must address not just access, but meaningful participation in AI development. The environmental impact of AI systems is also increasingly relevant. Large-scale models require significant computational resources, raising concerns about energy consumption and sustainability. Finally, there is a need to address cultural and linguistic inclusivity in AI systems. Many AI models are dominated by data from a limited set of languages and contexts, which can marginalize underrepresented communities. Addressing these cross-cutting issues will be essential to ensure that AI governance frameworks remain future-ready, inclusive, and globally relevant.

In Brunei and similar developing contexts, gaps in AI governance are already shaping both challenges and emerging opportunities, particularly within the education sector. One of the most significant challenges is the lack of clear regulatory frameworks and practical guidelines for AI adoption. Educators and institutions are increasingly experimenting with AI tools, yet there is limited guidance on issues such as data privacy, ethical use, and accountability. This creates uncertainty and slows down responsible implementation. A second challenge is capacity disparity. While AI technologies are advancing rapidly, local capacity in terms of expertise, infrastructure, and training remains limited. This risks widening the gap between countries that can fully leverage AI and those that are primarily consumers of external technologies. There are also concerns around trust and transparency. Without clear standards for explainability and oversight, stakeholders—including teachers, students, and parents—may hesitate to fully adopt AI-driven solutions, particularly in sensitive areas like assessment and personalized learning. However, these gaps also present important opportunities. As a smaller nation, Brunei has the potential to adopt agile, forward-looking governance frameworks that can be implemented more quickly than in larger systems. There is an opportunity to position education as a testbed for human-centered AI, where policies can be developed alongside real-world classroom applications. Furthermore, the growing awareness of governance challenges is encouraging collaboration between educators, policymakers, and technologists, creating space for inclusive, locally relevant solutions. Addressing these gaps effectively could enable countries like Brunei to move from being passive adopters to active contributors in shaping responsible AI use, particularly in education.

The Global Dialogue on AI Governance can play a pivotal role as a neutral, inclusive platform that bridges gaps between countries, sectors, and levels of AI readiness. First, it can facilitate the development of shared principles and interoperable governance frameworks, enabling countries to align on core standards such as safety, transparency, accountability, and human rights while still allowing flexibility for local adaptation. This is especially important in avoiding fragmented regulatory approaches that could hinder innovation and cross-border collaboration. Second, the Dialogue can act as a connector between stakeholders, bringing together governments, educators, industry leaders, researchers, and civil society. AI governance is inherently multidisciplinary, and meaningful cooperation requires continuous engagement across these groups rather than isolated efforts. Third, it can support capacity-building partnerships, particularly for developing and smaller nations. By enabling knowledge transfer, training programs, and access to tools, the Dialogue can help reduce global disparities in AI readiness and ensure more equitable participation. Fourth, the Dialogue can promote trust-building mechanisms, such as shared evaluation standards, transparency benchmarks, and collaborative pilot initiatives. Trust is essential for international cooperation, especially when AI systems operate across borders. Finally, it can serve as a platform for continuous learning and adaptation, ensuring that governance approaches evolve alongside rapid technological advancements, including emerging areas such as agentic AI systems. In essence, the Dialogue can move global AI governance from fragmented discussions to coordinated, inclusive, and action-oriented collaboration.

The AI Dialogue should build upon and connect with existing global and regional initiatives such as UNESCO's AI Ethics Recommendation, the OECD AI Principles, the Global Partnership on AI (GPAI), and emerging regulatory frameworks like the EU AI Act. It should also engage with industry-led and open initiatives, including efforts around open-source AI models and safety research communities. These initiatives have already established valuable foundations in areas such as ethical principles, policy guidance, and technical standards. However, they often operate in parallel rather than in a fully coordinated manner, and their implementation varies significantly across regions. The added value of the AI Dialogue lies in its ability to synthesize and operationalize these efforts into a more cohesive global approach. Rather than duplicating existing work, the Dialogue can act as a convergence platform, aligning different frameworks and identifying practical pathways for implementation, especially in underrepresented regions. Additionally, the Dialogue can bring stronger representation from education sectors, developing countries, and grassroots innovators, whose perspectives are often less visible in global policy discussions. This would ensure that governance frameworks are not only technically sound but also socially relevant and inclusive. Another key contribution would be the promotion of real-world pilot collaborations, where policies are tested in practical settings such as education, healthcare, or agriculture. This bridges the gap between high-level principles and actual deployment. Ultimately, the AI Dialogue can enhance global efforts by providing coordination, inclusivity, and actionable implementation pathways, ensuring that AI governance evolves in a connected and impactful way.

Different stakeholders can contribute to the AI Dialogue by bringing complementary perspectives grounded in real-world experience. Governments can provide policy direction and regulatory insights, while industry contributes technical expertise and implementation experience. Educators and researchers offer evidence-based insights and highlight practical impacts in sectors such as education. Civil society organizations ensure that ethical, social, and human rights considerations remain central. To maximize these contributions, the Dialogue should adopt a multi-layered structure. This could include: • Plenary sessions to establish shared priorities and global principles • Thematic working groups focused on key areas such as safety, capacity-building, and human rights • Regional roundtables to capture local contexts and challenges • Sector-specific tracks (e.g., education, healthcare, agriculture) to connect policy with real-world applications Importantly, the Dialogue should include mechanisms for continuous engagement, such as follow-up working groups or digital collaboration platforms, ensuring that discussions extend beyond a single event. A successful format should balance high-level policy dialogue with practical implementation insights, enabling stakeholders not only to share perspectives but also to co-develop actionable solutions.

Several important voices remain underrepresented in global AI governance discussions. First, educators and frontline practitioners are often overlooked, despite being directly responsible for implementing AI in real-world settings. Their insights are critical for understanding how policies translate into practice. Second, developing and smaller nations, including those in Southeast Asia, are underrepresented. These regions face unique challenges related to infrastructure, capacity, and resource constraints, which are not always reflected in global frameworks. Third, students and youth—who are both primary users and future developers of AI—are rarely included in governance discussions. Their perspectives are essential, particularly in education-focused applications. Fourth, non-technical communities, including parents, local communities, and marginalized groups, are often excluded due to technical complexity or lack of access. This can lead to governance approaches that do not fully reflect societal needs. To address this, the Dialogue should adopt inclusive participation mechanisms, such as: • Targeted invitations and representation quotas for underrepresented regions and sectors • Simplified communication formats to make discussions accessible to non-experts • Youth panels and practitioner-led sessions • Hybrid participation (virtual + in-person) to reduce access barriers Ensuring diverse participation will lead to more equitable, context-aware, and effective AI governance outcomes.

To foster meaningful and dynamic engagement, the AI Dialogue should move beyond traditional panel discussions and incorporate interactive, participatory formats. One effective approach is scenario-based workshops, where participants collaboratively explore real-world use cases—such as AI in education or healthcare—and identify governance challenges and solutions. This grounds discussions in practical realities. Another format is policy co-creation labs, where diverse stakeholders work together to draft actionable guidelines or frameworks during the Dialogue itself. This encourages ownership and accelerates implementation. Live demonstrations and case showcases can also be valuable, allowing innovators and practitioners to present working AI systems. This helps bridge the gap between policy and practice, making discussions more tangible. The Dialogue could also include multi-stakeholder simulations or role-playing exercises, where participants take on different perspectives (e.g., regulator, developer, educator) to better understand trade-offs and decision-making complexities. Additionally, digital collaboration platforms can enable real-time input, polling, and feedback, ensuring that even remote participants can actively contribute. Finally, incorporating youth-led sessions and innovation challenges can bring fresh perspectives and energy into the Dialogue. By combining these formats, the AI Dialogue can become more interactive, inclusive, and outcome-driven, ensuring that participants are not just passive listeners but active contributors to shaping AI governance.

Effective AI governance is already being advanced through a combination of policy frameworks, technical practices, and real-world implementation approaches. At the policy level, frameworks such as UNESCO's Recommendation on the Ethics of AI and the OECD AI Principles provide widely adopted guidance on human-centered, transparent, and accountable AI. More recently, regulatory approaches like the EU AI Act demonstrate how risk-based classification can translate ethical principles into enforceable requirements. From a practical perspective, risk-based governance models are particularly effective. By categorizing AI systems based on their level of impact (e.g., low-risk vs high-risk applications), stakeholders can apply proportionate safeguards without unnecessarily restricting innovation. In implementation, human-in-the-loop systems are a key good practice. Ensuring that AI supports, rather than replaces, human decision-making-especially in sectors like education-helps maintain accountability and trust. Another important approach is the use of transparency and explainability mechanisms, such as clear documentation of model behavior, data sources, and limitations. This allows users to better understand and evaluate AI outputs. Open ecosystems also play a role. Open-source models, shared datasets, and collaborative platforms enable wider participation, peer review, and faster innovation, while also supporting capacity-building in under-resourced regions. At the application level, AI-powered platforms in education that incorporate adaptive learning, accessibility features, and ethical safeguards demonstrate how governance principles can be embedded directly into system design. Overall, the most effective approaches combine policy guidance, technical safeguards, and real-world application, ensuring that AI governance is not only principled but also practical and scalable across different contexts.