University of Brighton, UK

A successful first Global Dialogue on AI Governance should deliver outcomes that move beyond consensus-building toward measurable and inclusive action. First, it should produce a clear and actionable Global AI Capacity and Equity Framework that identifies priority gaps in compute, data, skills, and institutional readiness, particularly in low and middle income countries, and outlines pathways for coordinated investment and support. Second, the Dialogue should result in a set of practical policy toolkits for high impact sectors such as health and public administration. These should translate existing principles into operational guidance, including risk classification, accountability mechanisms, and standards for transparency and human oversight. Third, it should establish a multi stakeholder implementation platform to facilitate ongoing collaboration between governments, private sector actors, civil society, academia, and international organizations. This platform should support knowledge sharing, technical assistance, and the scaling of context appropriate solutions. Fourth, the Dialogue should secure concrete and time bound commitments from key actors, including capacity building partnerships, data sharing initiatives, and responsible deployment practices. These commitments should be accompanied by mechanisms for tracking progress and ensuring accountability. Finally, success would require embedding equity and human rights as operational priorities, including an explicit focus on intersectionality to ensure that governance approaches address compounded and systemic harms.

• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight
• AI capacity-building

We have prioritised AI capacity building, transparency accountability and human oversight, protection and promotion of human rights, and the social economic ethical cultural linguistic and technical implications of AI, as these areas are critical to bridging the gap between high level principles and equitable real world implementation. Together they address both structural and operational challenges, ranging from global disparities in compute data and skills to the need for enforceable accountability mechanisms in high impact sectors such as health. This combination also enables a strong focus on intersectionality and data justice, recognising that AI systems can produce compounded harms across gender socioeconomic status geography and other axes of inequality. Prioritising these areas will support a more inclusive rights based and context sensitive approach to AI governance, while ensuring that capacity and accountability are strengthened in parallel, particularly in low and middle income settings.

Yes. While the listed themes are comprehensive, several cross-cutting and emerging issues require more explicit attention. First, the political economy of AI, particularly the concentration of compute infrastructure, data resources, and market power, remains insufficiently addressed. Without confronting these structural imbalances, efforts on capacity building and governance risk reinforcing existing dependencies rather than reducing them. Second, data justice and value distribution deserve clearer focus. Current governance discussions often emphasise data protection, but less attention is given to questions of ownership, benefit sharing, and community governance, especially for populations whose data is extracted without meaningful participation or return. Third, the environmental and resource footprint of AI is an emerging concern. The energy, water, and material demands of large-scale AI systems have implications for sustainability and may disproportionately affect resource-constrained settings. Fourth, the rise of generative AI introduces new systemic risks, including information integrity, manipulation, and fraud. These risks are evolving rapidly and require governance approaches that are adaptive and anticipatory rather than reactive. Finally, there is a need to explicitly centre intersectionality as a cross-cutting lens. AI-related harms are rarely experienced in isolation and often emerge at the intersection of age, gender, socioeconomic status, geography, and other structural factors. Without integrating this perspective, governance responses may overlook those most affected. Addressing these issues alongside the existing themes would strengthen the Dialogue's ability to move from broad principles to more equitable, context-sensitive, and forward-looking governance outcomes.

Governance gaps in AI capacity building, accountability, human rights, and the broader social and technical implications of AI are already having tangible effects in my country and sector, particularly in public health and digital service delivery. Limited access to advanced compute infrastructure, high quality datasets, and technical expertise continues to constrain the ability of institutions to develop and deploy contextually relevant AI solutions. This creates a growing dependency on externally developed systems, which may not adequately reflect local needs, languages, or health priorities. At the same time, the absence of clear and enforceable accountability frameworks poses risks in high impact settings such as healthcare. AI tools are increasingly being introduced without sufficient validation, transparency, or mechanisms for redress, raising concerns around safety, reliability, and trust. Human rights considerations are also becoming more prominent. Issues related to data privacy, consent, and potential discrimination are particularly relevant in diverse and resource constrained settings, where existing structural inequalities may be amplified by AI systems. These impacts are best understood through an intersectional lens, where overlapping factors such as socioeconomic status, geography, gender, and digital literacy interact to produce compounded vulnerabilities. Furthermore, the broader social and economic implications of AI, including risks of exclusion due to language barriers or digital literacy gaps, are shaping how different populations experience these technologies. These challenges often disproportionately affect older adults, rural populations, and those with limited access to digital infrastructure. Overall, while advances in AI present significant opportunities, the current governance gaps risk reinforcing existing inequities unless addressed through coordinated, context sensitive, and capacity focused approaches.

The AI Dialogue can play a critical role as a convening and coordinating platform that moves international cooperation from fragmented discussions toward more aligned, inclusive, and action-oriented governance. First, it can help build convergence across diverse governance approaches by facilitating structured exchange between countries at different levels of technological and regulatory maturity. This can support greater coherence while still allowing for contextual flexibility, particularly for low and middle income countries. Second, the Dialogue can strengthen cooperation on capacity building by identifying shared needs and enabling coordinated investments in compute infrastructure, data systems, and skills development. It can also serve as a platform to match technical expertise and resources with countries and sectors that require support. Third, it can promote the development of practical and interoperable governance tools, including risk assessment frameworks, accountability mechanisms, and standards for transparency and human oversight. By grounding these tools in real world use cases, the Dialogue can help translate principles into implementable practices. Fourth, the Dialogue can foster multi stakeholder partnerships that bring together governments, private sector actors, civil society, academia, and international organisations. These partnerships are essential for addressing cross border challenges such as data governance, safety risks, and the societal impacts of AI. Finally, the Dialogue can help centre equity and human rights within international cooperation by encouraging approaches that account for intersectional impacts and systemic inequalities. In doing so, it can ensure that global AI governance efforts are not only coordinated, but also fair, inclusive, and responsive to the needs of diverse populations. Overall, the AI Dialogue has the potential to act as a bridge between global ambition and local implementation, enabling more effective and cooperative AI governance worldwide.

The AI Dialogue should build upon a range of existing international initiatives and partnerships that have advanced principles, standards, and coordination on AI governance. These include normative frameworks such as those developed by the United Nations Educational, Scientific and Cultural Organization on AI ethics, the work of the Organisation for Economic Co-operation and Development on AI principles and policy observatories, and the technical standard-setting efforts of bodies such as the International Telecommunication Union. It should also connect with multi-stakeholder platforms like the Global Partnership on Artificial Intelligence, as well as emerging regional and national governance initiatives. In addition, the Dialogue can draw on sector-specific efforts, including those in health and digital public infrastructure, where AI deployment is rapidly expanding but governance remains uneven. Capacity-building initiatives led by international organisations and development partners also provide a strong foundation for scaling equitable access to AI resources and expertise. The added value of the AI Dialogue lies in its ability to act as a bridge across these fragmented efforts. While many existing initiatives focus on principles, standards, or specific sectors, there is a clear gap in coordination, implementation, and inclusivity. The Dialogue can provide a platform to align these efforts, reduce duplication, and promote interoperability across governance approaches. Importantly, it can elevate perspectives from low and middle income countries, ensuring that global governance frameworks reflect diverse contexts and priorities. It can also facilitate the translation of high-level principles into actionable tools, foster partnerships that link capacity needs with resources, and promote accountability through shared commitments and progress tracking. By connecting existing mechanisms while addressing gaps in coordination and implementation, the AI Dialogue can strengthen a more coherent, inclusive, and action-oriented global AI governance ecosystem.

Different stakeholders can contribute to the AI Dialogue through clearly defined and complementary roles. Member States can share policy experiences and align national frameworks with global norms. The private sector can provide technical expertise and ensure responsible deployment. Civil society can highlight lived experiences and intersectional harms, while academia and the technical community can contribute evidence and evaluation. International organisations can facilitate coordination and support capacity building, particularly in resource constrained settings. To ensure meaningful engagement, the Dialogue should move beyond plenaries and adopt a focused structure. Problem oriented working groups can address specific use cases such as healthcare or public services. Policy labs can translate principles and scientific insights into actionable tools. Commitment sessions can enable stakeholders to announce measurable actions, supported by follow up mechanisms. Such a format would support a more interactive, inclusive, and action oriented approach to AI governance.

Several voices and perspectives remain underrepresented in global discussions on AI governance. First, stakeholders from low and middle income countries, particularly from the Global South, are often not adequately represented in agenda setting and standard development processes. This limits the extent to which governance frameworks reflect diverse institutional capacities, development priorities, and local realities. Second, marginalised and vulnerable communities, including rural populations, older adults, persons with disabilities, and those with limited digital literacy, are rarely meaningfully included. Their experiences are critical for understanding how AI systems can produce intersectional and compounded harms in real world contexts. Third, non English speaking communities and those operating in low resource linguistic environments are often excluded, both in terms of participation and in the design of AI systems themselves. This creates barriers to access and representation. Fourth, frontline practitioners, such as healthcare workers, educators, and public sector implementers, are frequently overlooked despite being directly involved in the deployment and use of AI systems. To address these gaps, the AI Dialogue should adopt more inclusive participation mechanisms. This includes providing financial and logistical support for participation from underrepresented regions, enabling multilingual engagement, and creating dedicated spaces for community led inputs. It should also integrate participatory approaches, such as consultations and co design processes, that bring lived experiences into governance discussions. Ensuring diverse and intersectional representation will be essential for developing AI governance frameworks that are equitable, context sensitive, and responsive to those most affected.

To foster meaningful and dynamic engagement, the AI Dialogue should adopt formats that prioritise interaction, co-creation, and practical problem solving over one-way statements. First, problem driven working sessions can bring diverse stakeholders together around specific use cases such as healthcare or public service delivery. These sessions should focus on jointly identifying risks, gaps, and context specific solutions, producing tangible outputs rather than general discussions. Second, policy labs can translate high level principles and scientific insights into actionable governance tools. Participants can work in small, multidisciplinary groups to develop draft frameworks, risk assessment models, or accountability mechanisms that can be tested and refined. Third, scenario based simulations can be used to explore emerging risks, such as AI enabled misinformation or system failures in critical sectors. These formats allow stakeholders to engage with realistic situations, improving preparedness and shared understanding. Fourth, commitment sessions can create space for stakeholders to announce specific, time bound actions, such as capacity building initiatives or responsible deployment practices, supported by mechanisms for follow up and accountability. Fifth, participatory dialogues and community panels can ensure that lived experiences, particularly from underrepresented and marginalised groups, inform discussions. This is essential for capturing intersectional impacts that may not be visible in technical or policy debates.

Several existing policies, practices, and platforms offer concrete approaches to advancing effective AI governance. At the normative level, the Recommendation on the Ethics of Artificial Intelligence developed by United Nations Educational, Scientific and Cultural Organization provides a comprehensive global framework grounded in human rights, inclusion, and accountability. Similarly, the AI Principles of the Organisation for Economic Co-operation and Development have informed national strategies and promote values such as transparency, robustness, and fairness. At the regulatory level, the European Union Artificial Intelligence Act introduces a risk based approach that classifies AI systems according to their potential impact and establishes corresponding obligations. This offers a practical model for operationalising governance through enforceable requirements. In terms of tools and implementation practices, algorithmic impact assessments and audit frameworks are increasingly used to evaluate risks before deployment, particularly in public sector contexts. Regulatory sandboxes have also emerged as a useful mechanism, allowing innovators to test AI systems under oversight while informing adaptive regulation. Multi stakeholder platforms such as the Global Partnership on Artificial Intelligence support collaboration across governments, industry, and academia, while initiatives led by the International Telecommunication Union contribute to technical standard setting and global coordination. Finally, capacity building approaches, including national AI strategies and regional training initiatives, are essential for enabling equitable participation in AI development and governance. When combined, these examples demonstrate that effective AI governance requires alignment between principles, regulatory mechanisms, technical tools, and inclusive collaboration platforms.