University of Hertfordshire

In my view, the first Global Dialogue on AI Governance will be successful if it moves decisively beyond symbolic agreement and establishes durable foundations for translation, coordination, and trust across the multilateral system. Success should not be measured by the breadth of participation alone, but by whether the Dialogue demonstrably strengthens the United Nations' ability to govern artificial intelligence in a way that is adaptive, evidence based, and operationally meaningful, as envisaged in both the Pact for the Future and resolution A/RES/79/325. From the perspective of my doctoral research on risk appetite calibration in uncertain and high volatility contexts, a key outcome would be shared understanding among Member States and stakeholders about how acceptable AI risk differs by institutional capacity, societal trust, and stage of development. AI governance will fail if it assumes uniform readiness. A successful Dialogue should therefore surface practical mechanisms for calibrating risk, oversight, and accountability rather than defaulting to one size fits all principles. As a co author of two books on mixed reality leadership and AI governance, I also see success in whether the Dialogue enables leaders to govern across blended human–machine environments. This means embedding human oversight, decision authority, and responsibility clearly within AI enabled systems, rather than treating these as abstract ethical commitments. Finally, the Dialogue should deliver tangible pathways for capacity building and interoperability, particularly for developing countries and smaller actors, including small-medium enterprises (SMEs). If the first Dialogue produces a credible agenda that aligns scientific evidence from the Independent International Scientific Panel on AI with actionable governance support, it will have laid the groundwork for restoring trust in multilateral cooperation and positioning the UN as a convenor that is not only legitimate, but operationally effective.

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

From my perspective, the four thematic areas of transparency, accountability and human oversight; interoperability of governance approaches; AI capacity building; and the social, economic, ethical, cultural, linguistic and technical implications of AI most urgently require active engagement, as they collectively reflect a governance first and leadership aware approach to AI. My doctoral research and practitioner work focus on how risk, responsibility and decision authority are allocated when humans and AI systems operate together in conditions of uncertainty. Transparency, accountability and human oversight therefore form the primary anchor for my engagement, as they operationalise human in the loop governance and ensure that responsibility remains intelligible, contestable and grounded in human judgement, rather than abstracted into technical systems. Interoperability of governance approaches is equally critical because risk based AI governance cannot function in fragmented regulatory or institutional environments. My research sits at the intersection of agile practice, institutional design and multilateral coordination, and highlights that governance regimes must be able to "speak to one another" if they are to support responsible deployment across borders, sectors and post conflict or transitional contexts. Interoperability enables contextual flexibility while avoiding regulatory incoherence. AI capacity building is a necessary companion to these themes. Governance ambition without commensurate institutional capability risks deepening existing divides. My work recognises uneven readiness across institutions, sectors and regions, and positions capacity building as the practical mechanism that translates governance principles into implementation, particularly for SMEs and developing contexts. Finally, engagement with the broader social, economic, ethical, cultural, linguistic and technical implications of AI is essential for mixed reality leadership. Governing AI within lived social and organisational systems allows leadership, behaviour and socio technical complexity to be addressed holistically, rather than treated as secondary considerations. Together, these four themes reinforce a coherent, practical and system oriented contribution to the Global Dialogue.

From my perspective, a key cross-cutting issue not fully captured by the listed themes individually is the governance of decision authority and risk ownership across human-AI systems operating in conditions of uncertainty. While transparency, accountability and human oversight address essential safeguards, they do not on their own resolve how leadership judgement is exercised when responsibility is distributed across people, institutions and AI systems. My doctoral research highlights that risk does not reside solely in technology or regulation, but in how decisions are authorised, escalated and contested in mixed-reality environments. This issue cuts across oversight, capacity-building, interoperability and socio-technical implications, yet is not explicitly named. Similarly, risk-appetite calibration emerges as an integrative governance challenge rather than a standalone theme. Interoperability assumes governance regimes can align, but alignment requires shared understanding of acceptable risk across contexts with very different institutional maturity, trust levels and capabilities. Without explicit mechanisms for calibrating risk, interoperability risks becoming formal rather than functional. Another cross-cutting issue is institutional readiness for adaptive governance. Capacity-building is often treated as a technical or skills issue, yet my research shows it is equally a leadership and organisational design challenge. Institutions must be able not only to deploy AI responsibly, but to revise governance dynamically as conditions change-especially in post-conflict, crisis or transnational settings. Finally, governing AI within social and cultural systems introduces temporal and relational dimensions that cut across ethics, human rights and oversight. Leadership in mixed-reality contexts requires foresight, legitimacy and trust over time, not simply compliance with existing rules. Together, these cross-cutting issues reinforce the need for a governance-first, leadership-aware approach that treats AI governance as a living system of decision-making, rather than a static set of technical or normative controls.

Across the UK, Europe and multiple sectors, governance gaps and uneven advances in the four thematic areas I selected are already shaping how AI is adopted, trusted and contested in practice. In the area of transparency, accountability and human oversight, progress in regulatory frameworks has not consistently translated into clarity over decision authority in real‑world human–AI systems. In the UK and European public sector, organisations frequently struggle to determine who ultimately carries responsibility when AI informs decisions, particularly in complex, mixed‑reality leadership contexts where judgement is distributed across people and systems. This creates uncertainty that inhibits responsible deployment rather than enabling it. Challenges around interoperability of governance approaches are especially visible across Europe, where differing national interpretations of AI governance, procurement rules and risk management practices limit cross‑border collaboration and scaling. From my research perspective, risk appetite cannot be effectively calibrated when governance regimes cannot meaningfully align, leading to fragmentation that affects sectors such as healthcare, infrastructure, education and SMEs operating transnationally. In terms of AI capacity‑building, uneven institutional readiness remains a critical constraint. While policy ambition in the UK and EU is high, many organisations—particularly SMEs and public bodies—lack the leadership capability, skills and governance maturity to operationalise risk‑based AI governance. This gap reinforces inequalities between sectors and regions and slows responsible innovation. Finally, insufficient attention to the social, economic, ethical, cultural, linguistic and technical implications of AI continues to affect legitimacy and adoption. Governing AI as a socio‑technical system embedded in organisational cultures is still emerging, despite its importance for trust, workforce adaptation and leadership effectiveness. Taken together, these gaps underline the need for governance approaches that integrate oversight, interoperability, capability and socio‑technical leadership, rather than treating them as isolated challenges.

The AI Dialogue can play a critical role in advancing international cooperation by acting as a bridging mechanism between shared principles and operational governance practice, particularly across diverse institutional, cultural and developmental contexts. From my perspective, the Dialogue's greatest value lies in its ability to centre transparency, accountability and human oversight as a common governance language, ensuring that cooperation is grounded in clarity about decision authority, responsibility and human judgement in AI enabled systems. This is especially important for mixed reality leadership contexts, where governance failures often arise not from technological gaps but from ambiguity about who carries risk and accountability across borders. The Dialogue is also uniquely positioned to advance cooperation through interoperability of governance approaches. International coordination cannot succeed if governance regimes remain fragmented or incompatible. By fostering shared understandings of risk based governance and acceptable variance across contexts, the Dialogue can enable countries and sectors to align without imposing uniformity, supporting cooperation in transnational, post conflict and cross sector settings. Equally, meaningful cooperation depends on AI capacity building. The Dialogue can help shift cooperation away from compliance driven models toward capability oriented support, recognising uneven institutional maturity and enabling participation by SMEs, public sector actors and developing contexts. Without this, international cooperation risks reinforcing existing asymmetries. Finally, cooperation must be informed by the social, economic, ethical, cultural, linguistic and technical implications of AI, rather than treating AI governance as a purely regulatory exercise. The Dialogue can legitimise socio technical and leadership perspectives, creating space for shared learning about how AI reshapes organisations, work and trust. In doing so, it can reinforce a governance first, leadership aware approach that is both adaptable and operationally credible across international contexts.

In my view, the AI Dialogue should deliberately build upon and integrate across existing international initiatives such as the Global Digital Compact, the Independent International Scientific Panel on AI, UNESCO's Recommendation on the Ethics of AI, the OECD AI Principles, and regional governance efforts within Europe and the UK, while also connecting to practitioner led fora such as AI for Good and sector specific public private partnerships. These initiatives have established important normative foundations, evidence based analysis and sectoral experimentation, particularly in areas of transparency, human oversight and ethical safeguards. However, they often remain fragmented across policy, technical and operational domains. The unique added value of the AI Dialogue lies in its ability to integrate these efforts through a leadership and governance centred lens, rather than duplicating existing guidance. By foregrounding transparency, accountability and human oversight, the Dialogue can connect ethical and regulatory initiatives with concrete questions of decision authority, responsibility and escalation in real world human–AI systems—an area that remains insufficiently addressed across most existing mechanisms. Through a focus on interoperability of governance approaches, it can act as a coordination point where regional, national and sectoral frameworks are translated into compatible risk based practices that enable cooperation without forcing uniformity. The Dialogue can also strengthen and align existing capacity building initiatives by ensuring they address institutional readiness, leadership capability and governance maturity, not just technical skills—particularly for SMEs and developing contexts that are underrepresented in many current forums. Finally, by embedding discussion of the social, economic, cultural and organisational implications of AI, the Dialogue can provide connective tissue between policy, leadership practice and socio technical reality. In doing so, it can transform existing initiatives from parallel efforts into a more coherent, operationally credible system of international AI governance cooperation.

Different stakeholders can contribute most effectively as a multi-discipline approach to the AI Dialogue if the format and structure are designed to surface decision authority, governance practice and institutional readiness, rather than only normative positions or technical demonstrations. Governments and regulators should contribute by articulating how transparency, accountability and human oversight are enacted in practice, including where responsibilities are unclear or contested in human–AI systems. Private sector actors and SMEs can contribute grounded insights into how governance requirements are experienced operationally and where interoperability gaps between regulatory regimes create friction or risk. Academia and the scientific community, including those working at the intersection of leadership, governance and socio technical systems, should support the Dialogue by translating evidence into frameworks that help calibrate risk, responsibility and human judgement under uncertainty. Civil society and workforce representatives can contribute perspectives on trust, legitimacy and social impact, ensuring governance remains connected to lived experience. To enable this, the Dialogue would benefit from a layered structure. A high level plenary should establish shared principles around human oversight and accountability, while smaller, cross stakeholder working sessions should focus on specific governance challenges such as interoperability of approaches, capacity building pathways, and institutional readiness. These sessions should be problem oriented rather than position based, encouraging participants to examine concrete scenarios where governance breaks down. Capacity building should be embedded across the Dialogue, not treated as a separate track, with explicit attention to leadership capability and organisational maturity, particularly for SMEs and developing contexts. Finally, the Dialogue's structure should legitimise discussion of the social, economic, cultural and organisational implications of AI as governance issues in their own right. By combining strategic plenaries with practitioner led, scenario based exchanges and iterative follow up between sessions, the AI Dialogue can become an operational forum for international cooperation, rather than a purely consultative or advocacy driven space.

In my view, several voices and perspectives remain underrepresented in global discussions on AI governance, particularly those closest to decision making under operational uncertainty rather than policy formulation or technical design. These include public sector practitioners, frontline leaders, SMEs, and organisations in post conflict or transitional settings who must exercise judgement in human–AI systems with limited capacity, evolving mandates and real accountability exposure. Their absence affects meaningful progress on transparency, accountability and human oversight, as governance debates often describe what should be overseen without adequately capturing how responsibility, escalation and authority function in practice. There is also limited representation from actors working at the boundaries between governance regimes—cross border programmes, multinational SMEs and service ecosystems—where interoperability of governance approaches is not theoretical but a daily constraint. Without these voices, interoperability risks remaining a formal aspiration rather than an operational reality. Furthermore, capacity constrained institutions are frequently spoken about rather than with. Organisations lacking mature AI governance capabilities, particularly in developing contexts and smaller enterprises, rarely shape discussions on AI capacity building, even though their experience is essential for designing feasible, risk based governance pathways. Their inclusion would ground global cooperation in equity and realism rather than assumption. Finally, leadership perspectives that integrate the social, economic, cultural, linguistic and organisational implications of AI remain marginal. Mixed reality leadership scholars, organisational practitioners and workforce representatives bring insight into trust, legitimacy, behavioural change and socio technical complexity—dimensions critical to sustainable governance but often treated as secondary to technical or legal considerations. These gaps could be addressed through more deliberate stakeholder selection, scenario based dialogue formats, and structured participation mechanisms that value lived governance experience alongside technical and policy expertise. Embedding practitioners and leadership focused voices would strengthen the AI Dialogue's legitimacy and operational relevance across diverse contexts.

Innovative engagement formats for the AI Dialogue should prioritise decision making realism, cross system learning and leadership practice, rather than relying solely on plenary discussion or scripted panel exchanges. From my perspective, formats that surface how transparency, accountability and human oversight operate in practice are essential. Scenario based simulations or "governance design labs" could enable participants to work through concrete human–AI decision scenarios, clarifying how authority, escalation and responsibility function under uncertainty. This would directly support mixed reality leadership by making human in the loop governance tangible rather than abstract. To advance interoperability of governance approaches, the Dialogue could employ cross jurisdictional case mapping sessions, where participants from different regions and sectors collaboratively trace how governance frameworks intersect, conflict or align across borders. Such formats would allow diverse regimes to "speak to each other" and encourage practical convergence without imposing uniformity, particularly valuable in post conflict and transnational contexts. Meaningful engagement on AI capacity building would benefit from tiered participation formats that recognise uneven institutional maturity. Peer learning clinics or capability focused roundtables pairing high capacity institutions with SMEs and developing context actors could shift discussion from aspiration to implementation, ensuring governance ambition is matched by feasible pathways. Finally, to fully integrate the social, economic, cultural and organisational implications of AI, the Dialogue should incorporate facilitated reflection spaces that bring together leaders, workforce representatives and socio technical experts. Narrative based exchanges, leadership retrospectives or longitudinal impact discussions can capture how AI reshapes trust, work and organisational behaviour over time. Together, these formats would foster dynamic, problem oriented engagement, reinforcing a governance first and leadership aware Dialogue that values lived experience, institutional design and operational credibility alongside policy and technical expertise.

Effective AI governance is most visible where policies, practices and platforms translate high level principles into clear decision authority, shared responsibility and practical capability. At organisational level, human in the loop governance practices-such as defined escalation thresholds, decision logs, accountable AI owners and override procedures-make transparency, accountability and human oversight operational in mixed reality leadership contexts via a book I have co-authored with two other authors that was published in Aug. 2025, and been used with international at the University of Hertfordshire Business School/as resource for SME participants on the UK Government funded 'Help to Grow' Management Programme. Risk based AI impact assessments and lifecycle assurance models used across the UK and Europe further support leaders in clarifying how judgement and responsibility are exercised when AI informs decisions. At a system level as I am conducting with my doctoral research studies looking at an human:AI risk integrated approach interface with hybrid environments. So, interoperability platforms are essential to reducing fragmentation between governance regimes. Initiatives such as the OECD AI Policy Observatory provide a shared evidence base and comparable governance frameworks, while the ITU AI Standards Exchange maps emerging AI standards to align technical, regulatory and sectoral approaches. In parallel, ISO/IEC JTC 1/SC 42 and European conformity assessment and regulatory sandbox mechanisms under the EU AI Act enable common vocabularies around risk, assurance and oversight without enforcing uniform regulation. Together, these allow governance regimes to 'talk to each other', supporting cross sector and cross border collaboration. In relation to AI capacity building, approaches that integrate governance training with organisational change-such as capability maturity models, leadership development programmes and peer learning networks-have proven more effective than technical upskilling alone, particularly for SMEs and public sector bodies. Finally, participatory design and socio technical evaluation methods that address the social, cultural and organisational implications of AI help sustain trust and legitimacy, reinforcing governance as a leadership practice rather than a compliance exercise.