KM Systemberatung

A successful Global Dialogue on AI Governance should lead to outcomes that bridge the gap between high-level principles and practical implementation. First, it should establish a shared understanding that AI governance is not only a regulatory issue, but an operational and organizational challenge. Clear guidance is needed on how governance can be implemented within real-world structures and decision-making processes. Second, the dialogue should promote practical frameworks and approaches that help organizations identify where AI is used, understand its impact, and ensure transparency and accountability. This includes making responsibility visible and enabling traceability of AI-supported decisions. Third, it should encourage collaboration across stakeholders, ensuring that governments, private sector actors, academia and civil society contribute not only perspectives, but also actionable solutions and best practices. Fourth, the dialogue should highlight the importance of continuous governance. AI systems evolve over time, and governance must accompany them throughout their lifecycle, rather than being limited to initial assessment or compliance checks. Finally, success would mean creating momentum for implementation. Beyond discussions, the dialogue should lead to concrete follow-up actions, pilot initiatives, and shared learning processes that translate governance principles into practice. AI governance will only be effective if it becomes part of how organizations operate, not just how they comply.

• Transparency, accountability, and human oversight
• Interoperability of governance approaches
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Safe, secure and trustworthy AI

My selection reflects a strong focus on the practical implementation of AI governance within organizations. Transparency, accountability and human oversight are central, as the increasing use of AI is fundamentally changing how decisions are prepared and made. It is essential that responsibility remains visible, traceable and clearly assigned, even when systems are involved in decision-making processes. The interoperability of governance approaches is equally important. Organizations are confronted with multiple frameworks, regulations and standards. Without alignment and integration, governance risks becoming fragmented and ineffective. Practical approaches are needed that can be embedded into existing structures and processes. The broader social, economic and ethical implications of AI are also critical. AI does not only affect technology, but reshapes organizations, decision-making, and societal structures. Governance must therefore consider these wider impacts to remain relevant and effective. Finally, safe, secure and trustworthy AI provides the necessary foundation. However, trust cannot be achieved through principles alone. It must be built through operational practices that ensure transparency, accountability and continuous oversight. Overall, my selection reflects the need to move from abstract principles towards actionable governance that works in real-world contexts.

Yes, one important cross-cutting issue that is not sufficiently captured by the listed themes is the gap between AI usage and AI governance in practice. Across many organizations, AI is already being used in a decentralized and pragmatic way. However, governance structures often lag behind or are not fully established. This creates a situation where AI systems influence decisions without sufficient transparency, accountability or organizational awareness. This gap is not only a regulatory issue, but an operational one. Many existing discussions focus on principles, frameworks or compliance requirements, but less on how governance can be embedded into everyday processes and decision-making structures. Another emerging issue is the shift in decision-making itself. AI does not only automate tasks, but increasingly shapes how decisions are prepared, prioritized and executed. This raises fundamental questions about responsibility, control and human oversight that go beyond traditional governance models. Finally, there is a need to move from static assessments towards continuous governance. AI systems evolve over time, and their impact can change depending on context and usage. Governance therefore needs to be dynamic, ongoing and integrated into the lifecycle of AI systems. Addressing these cross-cutting challenges is essential to ensure that AI governance becomes effective in practice, not only in theory.

In Switzerland and across many European contexts, AI adoption is progressing rapidly in both the public and private sectors. Organizations increasingly use AI in areas such as administration, healthcare, finance and decision support. However, governance structures often lag behind this development. One of the most significant challenges is the lack of transparency regarding where and how AI is actually used. In many cases, AI systems are introduced pragmatically at different levels of an organization, without a centralized overview or clear governance framework. This creates risks related to accountability, compliance and trust. Another challenge is the growing complexity of regulatory and governance requirements, particularly in relation to the EU AI Act. While the regulation provides an important framework, many organizations struggle to translate its requirements into practical implementation within their existing processes and structures. At the same time, this situation creates significant opportunities. There is a strong potential to develop governance approaches that are not only compliant, but also operational and embedded into daily decision-making. Organizations that succeed in making AI usage transparent and responsibility traceable can build trust internally and externally. In sectors such as public administration and healthcare, this is particularly relevant, as decisions supported by AI can have direct societal impact. Effective governance can therefore become a key differentiator. Overall, the current governance gap represents both a risk and an opportunity: a risk if it remains unaddressed, and an opportunity to shape how AI is responsibly integrated into organizations and society.

The AI Dialogue can play a crucial role in advancing international cooperation by bridging the gap between global principles and practical implementation. Today, many frameworks and regulatory approaches to AI governance exist at national and regional levels. However, they often remain fragmented and difficult to translate into operational practice. The AI Dialogue can act as a platform to align these approaches, promote interoperability, and foster a shared understanding of how governance can be effectively implemented across different contexts. A key contribution of the Dialogue would be to move beyond high-level discussions and support the exchange of practical experiences. Organizations, governments and other stakeholders face similar challenges when integrating AI into real-world processes. Sharing lessons learned, use cases and governance approaches can accelerate progress and reduce duplication of efforts. The Dialogue can also strengthen trust between stakeholders by creating a space for continuous, inclusive and transparent exchange. This is particularly important given the global nature of AI and its cross-border impact. Furthermore, it can help promote a shift towards continuous governance. AI systems evolve over time, and governance must accompany them throughout their lifecycle. International cooperation is essential to develop approaches that remain effective in dynamic environments. Ultimately, the AI Dialogue can serve as a catalyst to translate governance principles into actionable, scalable and internationally aligned practices.

The AI Dialogue should build upon existing international, regional and sector-specific initiatives, while strengthening their connection to practical implementation. Key reference points include regulatory frameworks such as the EU AI Act, as well as global initiatives like the OECD AI Principles, UNESCO's Recommendation on the Ethics of AI, and ongoing discussions within the G7, G20 and other multilateral forums. These initiatives provide important guidance on principles, risk classification and ethical standards. However, a common challenge across these efforts is the translation of high-level principles into operational practice within organizations. This is where the AI Dialogue can create significant added value. The Dialogue can act as a connecting layer between policy, regulation and real-world implementation. By bringing together stakeholders from government, private sector, academia and civil society, it can facilitate the exchange of practical approaches, governance models and lessons learned. In addition, the Dialogue can help promote interoperability between different governance frameworks. Organizations often operate across jurisdictions and face multiple, sometimes overlapping requirements. Greater alignment and mutual understanding between initiatives would reduce complexity and improve effectiveness. Another key contribution would be to emphasize continuous governance. Existing initiatives often focus on classification or initial assessment, while less attention is given to how governance accompanies AI systems over time. The added value of the AI Dialogue lies in its ability to connect existing efforts, close the gap between principles and practice, and foster scalable, real-world governance approaches.

Different stakeholders can contribute to the AI Dialogue by bringing their specific perspectives, experiences and practical challenges into a shared space of exchange. Governments can provide regulatory frameworks and policy direction. The private sector can contribute real-world implementation experience, including the challenges of integrating AI into operational processes. Academia can offer research-based insights, while civil society can highlight societal impacts and ethical considerations. To be effective, the AI Dialogue should combine these perspectives in a structured and outcome-oriented way. Rather than focusing only on high-level discussions, it should include formats that enable practical exchange, such as case-based sessions, workshops and collaborative working groups. A key recommendation is to create spaces where stakeholders can share concrete use cases, governance approaches and lessons learned. This would help bridge the gap between principles and implementation. In addition, the Dialogue should be designed as a continuous process, not a one-time event. Regular exchanges, iterative feedback loops and follow-up activities are essential to ensure that discussions lead to real progress. Finally, accessibility and inclusiveness are critical. Participation should be open, transparent and supported by formats that allow contributions from different regions, sectors and levels of expertise. The strength of the AI Dialogue lies in its ability to connect diverse perspectives and translate them into actionable insights.

Several important perspectives are currently underrepresented in global discussions on AI governance, particularly those of practitioners working within organizations where AI is already being used in everyday operations. While policy-makers, large technology companies and academic institutions are well represented, less attention is often given to the experiences of organizations in sectors such as public administration, healthcare, education and small and medium-sized enterprises. These actors are directly confronted with the practical challenges of implementing AI, often without fully developed governance structures. Another underrepresented perspective is that of individuals responsible for operational decision-making. AI increasingly influences how decisions are prepared and executed, yet the voices of those managing these processes are rarely included in governance discussions. To address this, the AI Dialogue should actively include practitioners from different sectors and organizational levels. This could be achieved through targeted invitations, sector-specific working groups and platforms for sharing practical experiences. In addition, contributions should not be limited to formal statements. Interactive formats, case studies and real-world examples can help ensure that diverse perspectives are meaningfully integrated. Including these voices is essential to ensure that AI governance reflects not only theoretical considerations, but also the realities of how AI is used in practice.

Innovative engagement formats should focus on enabling practical exchange, co-creation and continuous interaction between stakeholders. One effective approach would be case-based dialogue formats, where participants present real-world use cases of AI implementation, including challenges, risks and governance solutions. This allows discussions to move beyond abstract principles and focus on concrete experiences. Another valuable format would be collaborative working sessions or "governance labs", where stakeholders from different sectors jointly develop approaches to specific governance challenges. These sessions could produce tangible outputs such as frameworks, guidelines or best practices. Interactive formats such as moderated roundtables and small-group discussions can further enhance meaningful exchange by allowing participants to engage more deeply and contribute actively, rather than remaining passive listeners. Digital and hybrid formats should also play a key role. Online platforms can enable continuous engagement beyond physical events, allowing stakeholders to share insights, comment on developments and collaborate over time. In addition, structured feedback loops are essential. Insights from discussions should be documented, shared and revisited in subsequent sessions to ensure continuity and progress. Finally, the inclusion of real-world perspectives is critical. Formats that integrate practitioners, decision-makers and operational stakeholders can help ensure that discussions remain grounded in practical realities. Innovative engagement formats should ultimately create a space where diverse perspectives lead to actionable outcomes and sustained collaboration.

Several existing policies, practices and approaches already contribute to effective AI governance, but their impact depends largely on how they are implemented in practice. Regulatory frameworks such as the EU AI Act provide an important structure for risk classification, accountability and compliance. Similarly, international principles like the OECD AI Principles and UNESCO's Recommendation on the Ethics of AI establish a shared foundation for trustworthy AI. In practice, however, organizations often struggle to translate these high-level frameworks into operational processes. Effective governance therefore requires complementary approaches that are embedded into daily workflows. One promising practice is the establishment of internal AI governance frameworks that create transparency about where AI is used, define clear responsibilities, and enable traceability of AI-supported decisions. This includes maintaining inventories of AI systems, documenting use cases, and integrating governance into existing decision-making structures. Another important approach is the concept of continuous governance. Rather than relying on one-time assessments, organizations need mechanisms to monitor, review and adapt AI systems over time as their context and impact evolve. Cross-functional collaboration is also essential. Bringing together technical, legal, operational and ethical perspectives within organizations helps ensure that governance is both comprehensive and practical. Finally, platforms that enable the sharing of best practices, use cases and governance models across sectors can significantly accelerate learning and implementation. Effective AI governance emerges where principles, regulation and operational practice are successfully connected.