BSI

As AI governance moves from rule setting to delivery, more effective connections are now needed between UN initiatives and the collaboration of public and private sector national bodies that deliver standards, measurement, assurance and market surveillance on the ground, together called the Quality Infrastructure (QI) system. The Dialogue is an opportunity to effectively connect high level negotiations and policy ambitions on AI governance , to the national, regional and international institutions that deliver governance services on the ground. The system of Quality Infrastructure (QI), made up of standards, measurement, assurance (conformity assessment and accreditation), and market surveillance, is the backbone of governance and interoperability across economies. The QI system operates somewhat under the radar, but is utilized by policymakers, businesses and consumers extensively. For example, businesses use certification (E.g. Quality Management or Health and Safety) to demonstrate their compliance with regulations or to demonstrate their competitive edge within the market to consumers. Similarly, policymakers utilize the system to provide their national markets access to global customers (e.g. through aligning policies with international standards) or to ensure consumers are safe through product testing and recalls. These QI services are usually delivered by designated national QI bodies on standards, measurement, accreditation and market surveillance.

• Safe, secure and trustworthy AI
• Interoperability of governance approaches
• Transparency, accountability, and human oversight
• Protection and promotion of human rights

Connecting effectively to existing governance ideas and mechanisms to reduce fragmentation is a key priority. The market is moving from principles based to codifying governance, however this is still fragmented. Pockets of established international principles such as OECD's AI principles and Due Diligence guidance, have led the way and are being combined with companies developing their own frameworks, such as Google's SAIF. The international standard system, such as ISO and IEC, can support market and UN efforts to develop consistency and reduce fragmentation by bringing together stakeholders from around the world, including business, International Organisations, and policymakers, who all hold an interest in AI Governance. By bringing together stakeholders, building consensus and codify best practice there is an opportunity to build global aligned frameworks on AI governance. International cooperation mechanisms in standards and assurance bodies, like ISO, IEC, BIPM and GLOBAL ACI, can enable the on-the-ground interoperability that high-level policy ambitions set out. For example, mutual recognition agreements across national accreditation bodies can enable AI-related assurance to scale across borders and avoid duplication. In addition, exploring how standards and assurance can enable more adaptive approaches that compliment regulation, including enabling faster paced changes in governance mechanisms to keep pace with technological change. It is also important to note that not all countries have equally resourced bodies delivering standards and assurance; in global supply chains and across global use of AI, there is an urgent need to build capacity in low- and middle-income countries within standards, assurance and measurement bodies to enable robust global AI governance. International Organisations such as World Bank, OECD, UNIDO and UNECE have made strides in this area, and we see value in the Dialogue joining up with these discussions to combine international resources in capacity building.

We would recommend a discussion on the best approach to utilizing existing governance frameworks to help deliver the ambitions of the global dialogue. Recognizing that existing Quality Infrastructure Systems can support in delivering interoperability, building consensus and reducing fragmentation. Building on the points highlighted in Question 1, we feel connecting the high-level policy ambition at UN and other International Organisations to the National QI systems would help operationalise AI governance in the longer term. We would encourage participation within this theme of international standards, measurement and accreditation organizations (e.g. ISO, IAF, BIPM and others), as well as interested National bodies, and particularly national QI bodies from low to middle income countries. BSI stands ready to support connecting the UN with the international institutions that are building the AI governance Frameworks and to explore opportunities for collaboration.

At international level, cooperation across Standards Development Organizations helps secure a strong framework of best practice that is built on global consensus. An example of this is the partnership of the World Standards Cooperation that looks at cooperation across standards between The International Electrotechnical Commission (IEC), the International Organisation for Standardization (ISO), and the International Telecommunication Union (ITU). The partnership has a focus on AI standardization and at their latest global conference in Seoul, Republic of Korea, they issued a joint commitment to advance the well-being of humanity through AI standards. An example of a global best practice that has joint SDO support is the ISO/IEC Management System Standard on AI, ISO/IEC 42001. The standard is a first-in-kind AI management system standard designed to enable the safe, secure and responsible use of Artificial Intelligence across society. The international standard is intended to assist organizations in responsibly using AI, addressing considerations like non-transparent automatic decision-making, the utilization of machine learning instead of human-coded logic for system design, and continuous learning. At national level the UK has developed the AI Standards Hub. Dedicated to knowledge sharing, community and capacity building, and strategic research, the Hub seeks to bring together industry, government, regulators, consumers and civil society, and academia with a view to facilitate the assessment and use of relevant published standards. The Hub acts as a resource for developing knowledge and skills needed to effectively engage with the field of AI standardization. E-learning materials and in-person training events aim to equip stakeholders to actively contribute to the development of standards and to use and interpret standards that have been published. The value of the Hub to elevating best practice in the UK has been recognised by international counterparts who have developed Hubs also to bring together the global community and share learning.

We would welcome the Global Dialogue to leverage the strength of the Quality Infrastructure system to operationalize the ambitions to scale globally aligned and credible AI governance. National QI bodies often have government recognition, and translate policy into action, such as ensuring products in the market are safe for consumers through use of standards, measurement, and assurance mechanisms. For example: • National Standards bodies turn high level principles such as "safe and trustworthy AI" into standards that developers and policymakers can apply. • Conformity assessment bodies such as testing and certification bodies provide independent assurance that systems meet those standards, supporting accountability and procurement decisions. • National Accreditation bodies assess that conformity assessment bodies are robust, ensuring that organizations providing certifications are credible, and reliable. • National Metrology Institutes ensure measurement frameworks are consistent and calibrated in a globally aligned way, ensuring data around the world are comparable. • Market surveillance authorities ensure that non compliant or unsafe products are identified and removed from the market. These national systems are internationally connected, through bodies such as International Organisation for Standardization, International Electrotechnical Commission (IEC), Global Accreditation Cooperation (Global ACI), Bureau International des Poids et Mesures (BIPM) where QI bodies represent their countries, allowing standards, measurements, and assurance to be recognised across borders. For instance, national standard bodies participate in international standards development such as in forums like ISO, IEC and other regional networks (e. g. COPANT in Pan-American region, CEN-CENELEC in Europe). Including the voice of these communities will help bring national perspectives on governance frameworks into the Dialogue discussions.

The National Quality Infrastructure is a long standing partnership between the public and private organizations that have developed trusted frameworks that underpin governance across sectors in nations, such as in medical devices, finance, health and safety, and sustainability. The system is also mobilising for AI governance, such as through the development of AI management system standards within ISO, and assurance frameworks in conformity assessment bodies. The strength of this system lies in its international corporation and consensus building, and the ability to bring stakeholders from around the world together through national government recognised bodies on standards and assurance, ultimately creating international rules and recognition of governance across borders. We would welcome the Global Dialogue to leverage the strength of the system to operationalize the ambitions to scale globally aligned and credible AI governance.