Blockchain&Climate Institute

Firstly, it is important to discuss and agree on a practical framework for global interoperability. While each jurisdiction can be expected to develop its own regulations, laying out coordination mechanisms early—such as mutual recognition of standards like ISO/IEC 42001 (Information technology — Artificial intelligence — Management system) and related ISO/IEC JTC 1/SC 42 deliverables—is paramount to the long-term success of global governance. This also means that the question of stakeholders needs to be addressed, and we believe that non-state actors deserve a "seat at the table," including established NGOs (particularly those with UN advisory status and ISO liaison status) and academia. Secondly, we believe in the importance of recognising the AI Divide across the North-South or other developmental axis. This means that the Global Dialogue should address access to vital technologies, know-how, and standards-based capacity-building tools. Finally, irrespective of the feasibility of designing and implementing enforcement mechanisms, it is important to agree on the "red lines" that constitute high-risk AI developments and applications, supported by transparent risk-management frameworks such as ISO/IEC 23894.

• Safe, secure and trustworthy AI
• AI capacity-building
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

The selected priorities reflect a structured and forward-looking approach to AI, emphasising interoperability, security, capacity-building, and the importance of transparency and accountability. Ensuring safe, secure and trustworthy AI is foundational, as it underpins adoption and helps prevent systemic risks. BCI's contributions to ISO/IEC JTC 1/SC 42 underscore the value of standards such as ISO/IEC 42001:2023 (AI Management System standard), which provides organisations with a certifiable framework for establishing governance structures, risk assessment, and continuous improvement, and ISO/IEC 23894:2023 (Guidance on risk management for AI), which operationalises these principles across the AI lifecycle. Safety and security are also central to emerging global regulatory standards, reinforcing their importance in global approaches to AI. We also emphasise transparency, accountability, and human oversight, as these are critical for the responsible use of AI in practice. Effective oversight mechanisms, explainability standards, and human review-supported by standards such as ISO/IEC 42005 (AI system impact assessment)-are necessary to bridge the gap between high-level principles and enforceable obligations, and are a prerequisite for the clear allocation of responsibility. In a fragmented regulatory landscape, interoperability of governance approaches becomes essential. Diverging national and regional frameworks create compliance burdens, legal uncertainty, and barriers to innovation. Promoting convergence through common standards, mutual recognition, and international cooperation-such as those developed under ISO/IEC JTC 1/SC 42 and ITU AI standards-can facilitate responsible cross-border AI development and deployment. In this context, technical and operational standards play a critical role by translating general principles into practical, implementable requirements. They support interoperability, reduce compliance complexity, and enhance trust, while enabling organisations, including smaller actors, to operationalise AI governance in a consistent and effective manner. BCI's official role in ISO/IEC AI standards further equips us to address the convergence of these technologies for verifiable climate action. Finally, AI capacity-building helps address disparities in access to skills, infrastructure, and resources, particularly in developing countries. Implementing AI solutions requires specialised technical expertise, including data scientists, software developers, and cybersecurity experts. In many areas, there may be a shortage of these skilled professionals, making it difficult to develop and implement these technologies effectively. Limited access to education and training in advanced technologies can delay the adoption and effective use of these tools. Strengthening capacity is essential to enable inclusive participation in the AI ecosystem and to ensure that the benefits of AI are widely shared. Standards such as ISO/IEC 42001 serve as practical tools for capacity-building by providing clear, implementable requirements that can be adopted globally, including through training and certification programmes.

The environmental and climate impact of AI systems is becoming increasingly significant. The growing energy consumption associated with large-scale models and data infrastructure raises important sustainability considerations, which should be addressed together with broader climate and sustainable development objectives. ITU's work on AI environmental efficiency standards (e.g., through the Focus Group on AI for Environmental Efficiency) and emerging ISO efforts provide valuable methodologies for measurement and mitigation that the Dialogue should integrate. The convergence of AI with other emerging technologies, such as distributed ledger technology (DLT), Internet of Things (IoT), and cloud computing-areas in which BCI actively develops standards-introduces new layers of complexity. These interactions can amplify both opportunities (e.g., transparent, verifiable ESG data for carbon markets) and risks, requiring more integrated and coordinated governance approaches. Data governance and access remain foundational challenges. Issues relating to data quality, data collection, availability, and cross-border data flows directly affect the performance, fairness, and reliability of AI systems. Standards such as the ISO/IEC 5259 series on data quality for machine learning are essential here. In addition, ensuring representative and unbiased datasets is essential to mitigate risks of discrimination and unequal outcomes. The increasing concentration of AI capabilities, particularly in relation to access to advanced computing infrastructure and large-scale models, raises concerns regarding competition, innovation, and equitable participation in the global AI ecosystem. This concentration may also create dependencies on a limited number of providers, with implications for resilience and digital sovereignty. Addressing these cross-cutting issues will be important to ensure that AI governance frameworks remain comprehensive, forward-looking, and responsive to technological and societal developments.

The AI Dialogue has the potential to bridge the inclusive priorities of the India AI Impact Summit and the technical safety standards of the ITU while promoting the adoption and harmonisation of ISO standards such as ISO/IEC 42001 and 23894. Hence, it should take this opportunity to serve as a high-level coordination platform that enables alignment of development priorities with ITU and ISO standards. Finally, the AI Dialogue at the UN level should consider the results and outcomes of the OECD AI Policy Observatory and the Hiroshima AI Process (G7).

Different stakeholders can contribute to the AI Dialogue, offering unique as well as complementary perspectives and insights. Actors with a global and cross-sectoral perspective—such as ISO liaison organisations like BCI—can play an important role in connecting policy discussions with technical developments and real-world implementation challenges. Governments can provide policy direction. International organisations can facilitate coordination and alignment, bringing broad perspectives and practical experience. The private sector, including both large companies and smaller innovators, contributes technical knowledge and operational insights, while academia and research institutions provide independent, evidence-based analysis. Broader stakeholder engagement helps ensure that societal impacts and diverse user perspectives are adequately reflected. To ensure meaningful participation, the Dialogue should adopt an inclusive, multi-stakeholder format with balanced geographic representation, particularly from developing countries. Enabling participation through capacity-building, funding, and access to expertise is essential to ensure that a wide range of actors and stakeholders can contribute effectively. The Dialogue would also benefit from open consultation processes, such as written submissions and stakeholder feedback mechanisms, as well as ongoing engagement channels between meetings. Clear and actionable outputs, including summaries, best practices, policy-oriented recommendations, and references to international standards, are important to ensure that discussions translate into practical and widely applicable outcomes.