NextGenMedics

The first Global Dialogue on AI Governance will only be meaningful if it moves beyond principles into enforceable, context-aware commitments. First, success would mean establishing a global baseline for high-risk AI systems, particularly in healthcare, that requires testing not only for accuracy, but for performance under real-world conditions such as incomplete data, unstable infrastructure, and limited technical capacity. AI systems that cannot function safely under these conditions should not be deployed in such environments. Second, the Dialogue should produce a framework for context-specific governance, recognizing that AI does not operate in a vacuum. Regulatory guidance must account for differences in data quality, infrastructure reliability, and workforce training across regions, especially in emerging health systems. Third, success would include a commitment to equitable data representation, with concrete steps toward developing and funding datasets from underrepresented regions. Without this, global AI systems will continue to embed systemic bias while appearing technically sound. Finally, the Dialogue must define clear accountability pathways. When AI systems fail in critical sectors like healthcare, there must be clarity on responsibility—whether it lies with developers, deployers, or institutions. From my perspective, informed by engagement in both healthcare systems and AI-focused policy discussions, success is not the production of another framework, but the establishment of standards that can withstand the realities of implementation across diverse global contexts.

• Transparency, accountability, and human oversight
• AI capacity-building
• Safe, secure and trustworthy AI
• Interoperability of governance approaches

I selected these thematic areas because they reflect both the challenges I have encountered and the work I have actively undertaken at the intersection of healthcare, AI, and policy. Safe, secure and trustworthy AI is critical because AI systems often assume stable infrastructure, complete datasets, and trained personnel conditions rarely met in many healthcare settings. Through designing AI triage tools and engaging directly with clinical workflows in Nigerian laboratories, I have seen how AI can fail in real-world environments, posing tangible risks to patient safety. AI capacity-building is essential to ensure that healthcare workers, policymakers, and system operators can safely deploy, interpret, and oversee AI tools. I have contributed to this by training peers and emerging healthcare professionals in AI concepts and practical applications, helping them understand both capabilities and limitations. Interoperability of governance approaches matters because AI policies must translate across diverse legal, cultural, and healthcare contexts. I have actively worked toward this by exploring frameworks that align technical AI system design with policy and operational realities, fostering solutions that bridge gaps between global standards and local implementation. Finally, transparency, accountability, and human oversight are non-negotiable. My participation in UN policy dialogues, including the High-Level Political Forum side event, has reinforced the importance of clear human-in-the-loop mechanisms, auditability, and accountability pathways. Engaging directly in these discussions has strengthened my understanding of how governance frameworks must be practically enforceable. Collectively, these priorities reflect both my hands-on experience and active engagement: bridging technical design, operational realities, and policy to ensure AI can be deployed safely, equitably, and effectively.

While the thematic areas identified by the General Assembly capture important dimensions of AI governance, several cross-cutting issues remain underemphasized, particularly in healthcare and other high-stakes sectors. Context-sensitive validation is critical. Most AI governance frameworks assume standardized, high-quality datasets and stable infrastructure, which are rarely present in low-resource environments. Without rigorous testing under conditions of incomplete data, unreliable power, and limited technical capacity, AI failures are inevitable-and the consequences are real: patients may receive misdiagnoses, clinicians may make incorrect treatment decisions, and health systems bear the operational burden. Accountability and liability are also insufficiently addressed. When AI systems err, the question of who bears the brunt-developers, deploying institutions, or front-line healthcare workers-is often unclear. Effective governance must define responsibility and establish enforceable human-in-the-loop oversight to ensure errors do not disproportionately harm vulnerable populations. Local data representation remains another gap. Global AI systems are frequently trained on datasets from high-income countries, which can embed biases and reduce effectiveness in regions like sub-Saharan Africa. Governance frameworks should mandate inclusive data collection, regional validation, and equitable access to AI benefits. Finally, integration with operational workflows and continuous monitoring is often neglected. AI systems must interact seamlessly with clinical procedures, laboratory workflows, and human decision-making, while ongoing auditing ensures reliability, fairness, and adaptability over time. Addressing these issues requires governance that spans technical, operational, and policy domains. My experience designing AI triage tools, training healthcare workers, and engaging in UN policy dialogues highlights that without concrete accountability mechanisms, AI failures can disproportionately harm those already at risk, undermining the promise of technology in real-world healthcare settings.

In Nigeria and similar low-resource settings, governance gaps in AI are already producing tangible challenges, particularly in healthcare. Many AI systems deployed or proposed in clinical settings are designed for high-resource environments, assuming reliable data, stable infrastructure, and trained personnel. In practice, incomplete laboratory datasets, intermittent power, and limited technical literacy among healthcare staff compromise the safety, accuracy, and reliability of AI-assisted decision-making. A critical gap lies in accountability and oversight. When AI systems produce errors—such as misclassifying lab results or mis-prioritizing patient triage—the responsibility often falls ambiguously on frontline staff or under-resourced institutions rather than system designers or policymakers. This creates not only clinical risk but also legal and ethical uncertainty. Data representation and equity present another pressing challenge. Global AI models are rarely trained on datasets reflecting local disease prevalence, genetic diversity, or healthcare workflows. As a result, AI outputs may be biased or less effective, limiting the potential benefits for patients in the region. Despite these challenges, there are notable opportunities. Capacity-building initiatives—such as training healthcare professionals in AI literacy and human-in-the-loop practices—can strengthen safe deployment. Efforts toward interoperability of governance frameworks can help align local regulations with emerging global standards, enabling responsible adoption without compromising context-specific realities. Furthermore, active engagement in policy dialogues and pilot projects provides a platform to shape governance approaches that are both globally informed and locally actionable. Addressing these gaps offers the dual benefit of improving AI safety and equity while positioning Nigeria and the region as contributors to global AI governance solutions. My experience designing AI triage tools, training healthcare personnel, and participating in UN policy forums underscores both the urgency and feasibility of bridging these gaps to ensure AI delivers tangible, equitable outcomes in real-world healthcare settings.

The AI Dialogue can play a transformative role in advancing international cooperation by bridging the persistent gap between high-level AI governance principles and real-world implementation, particularly in healthcare and other high-risk sectors. Current frameworks often assume ideal conditions—complete datasets, robust infrastructure, and highly trained personnel—which are rarely met in many countries. By providing a platform where policymakers, technical experts, and civil society practitioners converge, the Dialogue can ensure that governance approaches are informed by practical realities. One critical role is in defining global standards that are context-sensitive. The Dialogue can facilitate agreements on minimum requirements for safety, accountability, and human oversight that account for diverse operational environments, helping countries adapt global guidelines to local contexts. This includes clarity on liability when AI systems fail, mechanisms for auditing AI in practice, and protocols for integrating AI into existing workflows. The Dialogue can also accelerate capacity-building and knowledge sharing, fostering international collaboration to train technical personnel, clinicians, and policymakers on AI literacy, ethical deployment, and risk management. By enabling structured exchange of case studies and lessons learned, stakeholders can avoid repeating failures and scale effective practices more efficiently. Finally, the Dialogue can promote inclusive governance, ensuring that voices from underrepresented regions, sectors, and communities are heard in shaping AI policy. This is essential to prevent AI systems from embedding systemic inequities and to ensure that international standards reflect a diversity of experiences. From my perspective, grounded in both AI system design in healthcare and policy engagement at UN forums, the AI Dialogue can move beyond discussion to actionable, accountable, and globally relevant governance frameworks that are practical, equitable, and responsive to real-world challenges.

There are several existing initiatives, partnerships, and mechanisms in Africa that the AI Dialogue should build upon to ensure that international AI governance engagement is inclusive, actionable, and locally relevant. In recent years, the African Union (AU) has declared AI a strategic priority and adopted the Continental Artificial Intelligence Strategy, emphasizing inclusive, ethical, and sustainable AI ecosystems aligned with Agenda 2063 and the Sustainable Development Goals. This strategy provides a foundational policy framework that the AI Dialogue can connect with to deepen cooperation on governance norms, regulatory alignment, and capacity building across the continent. At the regional level, Smart Africa's Africa AI Council — convening African heads of state, AU representatives, ITU officials, and private sector actors — is advancing strategic guidance on AI infrastructure, talent, data, market, and governance across member states. This council represents a structured mechanism for harmonizing continental AI policy that the AI Dialogue could leverage for broader multi‑stakeholder cooperation. Ongoing collaborations between UNESCO, the African Union, and Smart Africa have produced inclusive data governance consultations and initiatives aimed at supporting rights‑based AI frameworks tailored to African contexts. Such work underscores the importance of data sovereignty, ethical adoption, and institutional readiness, all of which are essential complements to global governance discussions. Additionally, high‑level dialogues such as the AI for Africa Conference (under the G20) have launched the AI for Africa Initiative, bringing together policymakers, funders, and practitioners to mobilize resources and support implementation of governance and capacity objectives. The AI Dialogue can add value by linking these regional frameworks and multi‑stakeholder mechanisms to global governance standards, helping translate high‑level principles into actionable agreements that are informed by Africa's lived realities. It can also promote resource sharing, technical cooperation, and inclusive representation, ensuring that governance outcomes reflect both global priorities and context‑specific needs.

Different stakeholders; governments, technical communities, civil society, academia, and international organizations can contribute to the AI Dialogue by bringing their unique perspectives, expertise, and operational insights to the table. Governments can share regulatory experiences, pilot policy frameworks, and identify priority sectors for AI deployment. Technical communities, including AI developers and system designers, can provide insights on feasibility, limitations, and design considerations of AI tools, ensuring that governance standards are grounded in practical realities. Civil society and youth-led organizations can highlight societal impacts, equity considerations, and the human consequences of AI adoption, particularly in low-resource contexts such as healthcare in Africa. Academia can contribute evidence-based research, best practices, and ethical analyses, while international organizations can facilitate cross-border cooperation, provide technical assistance, and support knowledge exchange. To maximize impact, the Dialogue should be highly structured yet participatory. I recommend a hybrid format combining plenaries, focused roundtables, and sectoral labs. Plenaries can address overarching governance themes and emerging issues, while roundtables allow deep dives into specific sectors such as healthcare, finance, and education—linking technical, ethical, and policy perspectives. Sectoral labs can facilitate hands-on problem solving, collaborative development of guidelines, and simulation exercises to test governance approaches in realistic scenarios. Further, stakeholder engagement should be iterative and inclusive. Pre-Dialogue consultations, live polling, and post-session working groups can capture a wide range of inputs, including from regions often underrepresented in AI policy discussions. Transparent reporting and follow-up mechanisms are essential to translate discussions into concrete policy recommendations and capacity-building initiatives. Drawing from my experience designing AI tools for healthcare, training peers, and engaging in UN policy dialogues, I believe a structured yet participatory format—emphasizing real-world challenges, multi-stakeholder knowledge sharing, and actionable outcomes—will make the AI Dialogue both practical and transformative for global AI governance.

Despite increasing attention to AI governance, several critical voices remain underrepresented in global discussions, particularly from regions, sectors, and communities most affected by real-world AI deployment. In Africa, for example, healthcare practitioners, laboratory professionals, and policy implementers often face AI tools daily but have limited opportunities to influence governance frameworks. Their perspectives are essential to ensure that AI systems are safe, contextually relevant, and operationally feasible in low-resource settings. Other underrepresented groups include youth-led organizations, local AI developers, and technical communities outside of high-income countries. These stakeholders can provide insights into cultural, linguistic, and infrastructural nuances that are often overlooked in global AI models and standards. Marginalized communities—rural populations, people with disabilities, and those with limited digital access—also rarely have direct representation, even though they frequently bear the brunt of AI errors or biases. To include these voices, the AI Dialogue could adopt regional consultations and sector-specific workshops ahead of the main sessions, ensuring that inputs from underrepresented communities inform the agenda. Structured mechanisms such as advisory boards with diverse representation, live feedback channels, and multi-lingual participation platforms can help capture a broad range of experiences. Partnerships with local universities, professional associations, and civil society networks can facilitate outreach and capacity building, enabling participants to engage effectively in governance discussions. Drawing from my experience designing AI triage tools, training healthcare personnel, and engaging in UN policy dialogues, I have observed that meaningful inclusion requires more than token representation. It requires mechanisms for ongoing engagement, empowerment, and accountability, ensuring that underrepresented voices not only contribute but actively shape policies that affect the safety, equity, and effectiveness of AI systems in their contexts.

To foster meaningful and dynamic engagement during the AI Dialogue, formats should combine interactivity, real-world problem solving, and multi-stakeholder collaboration. Traditional plenaries, while important for framing discussions, are insufficient on their own to capture diverse perspectives or test actionable solutions. Hybrid sectoral labs are particularly effective. These small, focused sessions allow participants from different sectors—technical, policy, civil society, and healthcare—to collaboratively explore governance challenges, simulate real-world AI deployment scenarios, and co-develop practical policy or operational guidelines. For example, in healthcare, labs could simulate AI triage tool deployment in a low-resource hospital, highlighting safety, accountability, and workflow integration challenges. Interactive scenario workshops can complement labs by presenting participants with evolving, high-stakes AI situations, prompting them to make decisions under constraints, assess risks, and negotiate responsibilities. This format encourages debate, critical thinking, and problem-solving while revealing governance gaps that might otherwise remain theoretical. Multi-lingual digital engagement platforms can broaden participation from underrepresented regions, enabling live polling, Q&A, and collaborative document editing. These tools allow voices from low-resource settings, youth organizations, and rural communities to influence the dialogue without requiring physical presence. Iterative feedback loops, including pre-Dialogue surveys and post-session working groups, can capture insights in real time and ensure that discussions lead to actionable recommendations rather than remaining abstract. Drawing from my experience designing AI tools for healthcare, training peers, and participating in UN policy events, I have found that engagement is maximized when participants are active problem solvers rather than passive listeners. By combining labs, scenario workshops, digital platforms, and iterative feedback, the AI Dialogue can generate not only ideas but tangible, implementable outcomes that reflect diverse perspectives and practical realities.

Several policies, practices, platforms, and approaches demonstrate effective AI governance and offer solutions to its challenges, particularly in high-stakes sectors such as healthcare. Policies and frameworks: The European Union's AI Act provides a tiered risk-based regulatory approach, mandating stricter requirements for high-risk AI systems, including healthcare applications. This framework emphasizes safety, transparency, human oversight, and accountability, offering a model for context-sensitive governance globally. Similarly, the African Union's Continental AI Strategy promotes ethical, inclusive, and sustainable AI deployment across member states, highlighting the importance of regional alignment and capacity building. Practices: Human-in-the-loop systems in AI-assisted diagnostics illustrate operationalizing accountability. By ensuring that healthcare professionals validate AI outputs, these practices reduce errors, clarify responsibility, and enhance trust. Continuous monitoring, auditing, and model retraining based on real-world data further strengthen system reliability. Platforms: Initiatives such as Smart Africa's Africa AI Council and the UNESCO AI Policy Observatory provide structured mechanisms for multi-stakeholder engagement, knowledge sharing, and policy benchmarking. These platforms enable countries and institutions to align governance approaches, share lessons learned, and develop context-specific regulations. Approaches: Participatory governance and inclusive design methods ensure that AI systems reflect local realities. For instance, training healthcare personnel, engaging civil society, and involving local AI developers in deployment decisions help mitigate bias, ensure interoperability, and build capacity. Scenario-based simulations and sectoral labs-methods I have applied in healthcare AI projects-allow stakeholders to test governance frameworks against realistic operational challenges before full-scale deployment. These examples collectively demonstrate that effective AI governance requires integrating policy, technical practices, operational oversight, and inclusive participation. Combining global frameworks with locally grounded, participatory approaches ensures that AI is deployed safely, equitably, and effectively, particularly in resource-constrained contexts where the consequences of failure are most significant.