SHER DeepAI

A successful first Global Dialogue on AI Governance, led by the United Nations, should turn shared principles into practical and coordinated action. An important outcome would be stronger alignment on risk-based governance, helping connect frameworks such as the EU AI Act and standards like ISO/IEC 23894. Even a common understanding of how to define and manage risks would be meaningful progress. It should also highlight practical ways to implement governance—including tools for explainability, auditing, and continuous monitoring—so organizations can apply these principles in real-world settings. The Dialogue should also ensure inclusive participation and capacity building, enabling countries and institutions with varying levels of AI maturity to engage meaningfully and benefit from shared knowledge. Together, these outcomes would support more trustworthy, risk-aware, and globally aligned AI governance.

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

Our selection reflects a focus on making AI governance practical, trustworthy, and globally accessible. For safe, secure and trustworthy AI, we see an urgent need to strengthen risk governance across the full AI lifecycle, ensuring systems remain reliable and aligned with their intended use. At SHER DeepAI, we contribute through explainable AI (XAI) technologies that support risk identification, monitoring, and mitigation. AI capacity-building is equally critical. As a university lecturer and strategist, I actively teach and promote principles of trustworthy and responsible AI, helping equip professionals and students with the knowledge needed to engage with AI systems effectively. Transparency, accountability, and human oversight are central to building trust. Our XAI tools make AI systems more understandable, auditable, and actionable, enabling meaningful oversight and compliance in real-world applications. Finally, for open-source software, open data and open AI models, we believe openness must be combined with responsible and risk-aware use. SHER DeepAI supports this by providing explainability solutions that make open models more transparent and safer to deploy. Together, these priorities reflect our commitment to translating governance principles into practical, scalable, and impactful solutions

what remains missing is a shared and common understanding of what constitutes trustworthy and responsible AI across countries and regulatory frameworks. Today's landscape is fragmented, with a patchwork of guidelines and regulations across jurisdictions . There is a clear need for a globally accepted AI governance codex-a common foundation that aligns principles, standards, and implementation approaches. Such alignment would reduce fragmentation and enable more consistent and effective governance worldwide. I would welcome the opportunity to actively contribute to this effort.

In Germany and across Europe, AI governance is progressing rapidly, particularly with the EU AI Act. This creates strong opportunities to build trustworthy, human-centric AI and position Europe as a leader in responsible AI. A key challenge, however, is the gap between regulation and practical implementation. Many organizations struggle to operationalize requirements such as risk management, transparency, and human oversight. This is especially difficult for smaller organizations. At the same time, there are clear opportunities. The focus on transparency and accountability is increasing demand for explainable AI solutions, where SHER DeepAI contributes by making systems more understandable and auditable. Capacity-building is also essential. As a lecturer, I see growing demand for education on trustworthy and responsible AI, supporting more informed adoption. Overall, Europe has a strong foundation, but success depends on practical implementation, skills development, and global alignment.

The AI Dialogue, led by the United Nations, can serve as a neutral platform to strengthen international cooperation on AI governance. Its key role is to bridge fragmented approaches, aligning frameworks such as the EU AI Act, international standards like ISO/IEC 23894, and global principles. A key contribution of the Dialogue would be to move toward a shared understanding of trustworthy and responsible AI, and support the development of a globally accepted AI governance codex—a common framework aligning principles and implementation across countries. It can also promote knowledge sharing and capacity building, ensuring inclusive participation. In this way, the Dialogue can help shift from discussion to coordinated, practical, and globally consistent AI governance.

There is a rich landscape of local and applied initiatives, particularly in regions like Munich, including appliedAI Initiative, Munich Center for Machine Learning, and AI+MUNICH. These ecosystems play a key role in translating governance principles into real-world implementation. While these efforts provide a strong foundation, they remain fragmented across regions and levels. The added value of the United Nations AI Dialogue is to connect these initiatives globally, foster alignment, and support practical implementation. It can help build a shared understanding of trustworthy AI and contribute to a globally accepted AI governance codex, enabling more consistent and effective governance worldwide.

The AI Dialogue should be multi-stakeholder and action-oriented, with stronger inclusion of the business community, including AI developers, companies using AI, and those translating policy into practice—such as consultants, auditors, and technical experts. Governments can provide policy direction, while industry brings real-world experience in developing and deploying AI. Consultants, auditors, standardization bodies, and technical experts are essential because they translate governance principles into practical requirements, assurance mechanisms, and best-practice recommendations. Academia should contribute research and capacity building; as a lecturer, I promote principles of trustworthy and responsible AI. Civil society and end-users should ensure inclusiveness, accountability, and societal relevance. In terms of structure, the Dialogue should combine plenary sessions with focused working groups that produce practical outputs, including guidelines, tools, best-practice recommendations, and a globally accepted AI governance

Response Several important voices remain underrepresented in global AI governance discussions. First, the business community beyond large technology companies—including SMEs, startups, and companies deploying AI in traditional industries—is often overlooked. These actors face practical challenges in implementing governance and should be included through industry working groups and real-world use cases. Second, the "implementation layer"—consultants, auditors, and technical experts who translate policy into operational and technical requirements—is rarely represented. Their involvement is essential to make governance actionable and auditable. Third, standardization and assurance bodies, such as ISO and IEEE, are critical for turning principles into interoperable standards and certification mechanisms, yet are not always sufficiently integrated. Fourth, educators and capacity-building institutions play a key role in shaping understanding of trustworthy AI. As a lecturer, I see the importance of including academic voices that bridge theory and practice.

Effective engagement should combine conferences, roundtables, and interdisciplinary working groups. Conferences can set direction, while roundtables enable focused exchange between stakeholders. Interdisciplinary working groups are key to developing practical outputs such as guidelines and best practices, involving policymakers, industry, and implementation experts. The Dialogue should use hybrid formats (online and offline) to ensure broad and inclusive participation. Including real-world case studies helps make discussions more actionable.

On a practical level, approaches such as explainable AI (XAI) offer concrete solutions by making AI systems more transparent, enabling auditing, and supporting compliance. At SHER DeepAI, we apply XAI to help organizations better understand, monitor, and manage AI systems in line with governance requirements. In addition, local ecosystems and platforms, such as appliedAI Initiative, demonstrate how governance principles can be translated into real-world implementation through collaboration between industry, academia, and policymakers.