Independent Expert

As the former Chief of the Global Knowledge Management Unit at the World Food Programme, I believe the Global Dialogue has a unique opportunity to address a major gap in today's AI governance landscape: the lack of a coherent approach to knowledge governance. AI systems are only as reliable, equitable, and safe as the knowledge ecosystems that feed them. By explicitly linking AI governance with knowledge governance, the Dialogue can help ensure that global standards address not only model behavior, but also the quality and stewardship of the knowledge inputs that shape AI outcomes. This is essential for building trust, reducing systemic risk, and ensuring that AI serves all communities—not just those whose knowledge is already digitized. To enable a substantive discussion in July, I suggest three thematic clusters: 1. Knowledge Integrity and Architecture: covering data quality, metadata, and context preservation 2. Knowledge Equity and Inclusion: ensuring diverse linguistic, cultural, and community knowledge is represented. 3. Human Expertise and Sensemaking: recognizing the role of human judgment and participatory knowledge practices in AI oversight. The Dialogue should prioritize formats like interactive sessions that allow Member States, practitioners, and communities to surface tacit knowledge and lived experience, and feedback loops that ensure the Scientific Panel's report is not just presented, but contextualized and translated into actionable policy guidance. Key actors can contribute in the following ways: • Member States can share national knowledge governance practices. • The private sector can provide insight into real world data pipelines. • Civil society and communities bring lived experience and perspectives on equity and rights. • Academia and the technical community contribute critical research on knowledge representation and model interpretability. • International Organisations can help harmonize standards and support capacity building. Together, these actors form a distributed knowledge ecosystem that the Dialogue should intentionally activate. Finally, as outputs, The Dialogue should aim to deliver: • A shared framework linking AI governance with knowledge governance; • Guidance on organizational knowledge architecture; and • A set of pilot initiatives or partnerships that demonstrate how improved knowledge governance leads to safer, more trustworthy AI—whether in public health, climate action, humanitarian response, or education. These outputs would illustrate both the opportunities and the risks, while giving Member States and organizations concrete tools to act. In closing: If we want trustworthy AI, we must invest in trustworthy knowledge. AI governance and knowledge governance must evolve together, or neither will succeed.

Interoperability of governance approaches;Social, economic, ethical, cultural, linguistic and technical implications of AI;AI capacity-building;Transparency, accountability, and human oversight;

With the rapid advancement of AI, the AI that we speak about today will not be the AI we speak about next week. In this environment, AI governance must not only be strong but also able to adapt. Capacity building and inclusivity of AII governance ensure that it remains sustainable and relevant, and keeping "humans in the loop" ensures critical quality control and enhances accountability.

AI governance cannot be separated from knowledge governance. AI systems are only as reliable, equitable, and intelligent as the knowledge ecosystems they draw from. If we ultimately want trustworthy AI, we must invest in trustworthy knowledge: well structured, well governed, diverse, and ethically stewarded. Knowledge architecture, knowledge equity, and human sensemaking must be core pillars of global AI governance and not afterthoughts.

AI is not just a "technology" or "digital" issue. Many times, AI discussions are happening without Knowledge Management practitioners. As AI systems are only as reliable, equitable, and intelligent as the knowledge ecosystems they draw from, it is critical that knowledge architecture, knowledge equity, and human sensemaking must be core pillars of global AI governance. Good practices and lessons learned must be shared and built upon. Local and indigenous knowledge must inform AI governance discussions. Knowledge and AI must go hand in hand for AI governance to be as effective as it needs to be to incite the change we want to see in public health, climate action, humanitarian response, education and other sectors.

The AI Dialogue can convene a wide range of stakeholders to share good practice and experience, which will advance cooperation. The 2016 JIU report on KM in the UN system called out the UN as a catalyst and a disseminator of knowledge. Knowledge management can be used as a tool for promoting collaborating across departments, across the UN system, and with multi-stakeholder interventions.

-Working with the UN System Staff College as the dedicated hub for knowledge management and capacity building institution in the UN system; develop new courses and enhance the UNSSC's existing courses on Knowledge Management with AI governance principles. - Enhance UN DESA Digital Learning Center's course on KM fundamentals

- Member States can share national knowledge governance practices. • The private sector can provide insight into real world data pipelines. • Civil society and communities bring lived experience and perspectives on equity and rights. • Academia and the technical community contribute critical research on knowledge representation and model interpretability. • International Organisations can help harmonize standards and support capacity building.