91Â鶹¹ú²ú¾«Æ·×ÔÅÄ

The joint ISO (International Organization for Standardization) and IEC (International Electrotechnical Commission) committee on Artificial Intelligence (AI), SC 42, is playing a crucial role in developing international standards to facilitate the responsible adoption of AI technologies. SC 42 has an ecosystem approach that takes into account regulatory, business, societal and ethical perspectives to develop comprehensive standards that can be applied universally.  

As a committee under the IEC and ISO, SC 42 is under the one country, one vote system. ISO/IEC standards are voluntary and because they are reached by consensus, they foster collaboration and cooperation among a broad range of stakeholders. The diversity and inclusivity of the ISO/IEC process means that we are aligned with the needs of stakeholders all over the world — Africa, Asia, the Americas, Europe and Oceania. More than 60 countries currently participate in the work of SC 42, more than a third of which are considered developing countries. We have people of all ages and half of the delegations that attend our plenary meetings are led by women.  

A key focus of SC 42’s work is on creating international standards that enable trustworthy digital transformation. The ecosystem and platform approach to developing standards supports innovation by providing users with building blocks based on industry-wide best practices. These can also be combined seamlessly with application domain standards. SC 42 has published foundational standards on concepts, terminology, governance and frameworks for AI using machine learning, for example. The ISO/IEC committee has also developed a series of foundational standards for big data. 

In terms of the data ecosystem, SC 42 has developed standards for managing and overseeing big data analytics, emphasizing an industry-first approach to international standards. A series of standards on AI data quality, including terminology, measures, guidelines, and governance frameworks, is in progress. 

SC 42’s goal is to enable broad responsible adoption of AI systems while simultaneously unlocking the beneficial innovative aspects that AI can bring to society. To this end, SC 42 has developed a suite of standards that address emerging issues, including explainability, controllability, transparency, robustness, bias, oversight of AI systems, application of risk management frameworks and functional safety in AI systems. SC 42 takes a complementary approach, leveraging existing IT standards and adapting them for AI-specific requirements. 

Collaborating with sister IEC and ISO committees, SC 42 has developed standards on the governance implications of AI, offering guidance on defining responsibilities and assigning accountability. The committee explores computational aspects related to knowledge management and AI system performance, publishing overviews and methodologies. 

In collaboration with the ISO/IEC committee for software and systems engineering, SC 42 is addressing AI testing needs, providing guidelines for the testing of AI systems and verification/validation analysis. Elsewhere, SC 42 collaborates with the ISO’s committee on AI-enabled health informatics, promoting best practices, identifying use of AI concepts for standards development, and referencing relevant use cases. It also works with the ISO committee on language and terminology on natural language processing and with the IEC committee for industrial-process measurement, control and automation on functional safety.  

Sustainability and beneficial AI are additional areas of focus. SC 42 is developing standards to assess the environmental sustainability aspects of AI systems and studying beneficial AI outcomes, aiming to align with the UN Sustainable Development Goals (SDGs). The committee's work supports 12 of the 17 SDGs directly and all of the SDGs when SC 42’s horizontal standards are used in vertical domains across IEC and ISO committees.  

Ethical considerations and societal needs such as sustainability, transparency and trustworthiness are an integral part of all SC 42 deliverables. SC 42 has supported this aspect of its work with a standalone publication on AI ethics. SC 42 is currently working on a new technical specification offering guidance on addressing societal concerns and ethical considerations. SC 42 also recognizes the fundamental importance of collaboration and have built an extensive network of over 50 partnerships with organizations like UNESCO, OECD, the World Trade Organization (WTO) and the European Commission. 

Placing societal needs such as sustainability, transparency, trustworthiness and ethical considerations at the heart of our work can help unlock AI's full potential. It ensures responsible adoption while simultaneously unlocking the societal benefits that AI technology and innovation can bring globally. 

A good example is the recently published standard known as . It helps organizations of all types and sizes to develop and use AI systems responsibly by giving guidance on establishing, implementing, maintaining and continually improving an AI management system. The development of management system standards (MSS) such as ISO/IEC 42001 is a unique aspect of SC 42's approach providing a platform for third-party certification. This gives additional reassurance to stakeholders that the requirements have been implemented correctly. SC 42 is also planning standards for AI system impact assessment and third-party audit and certification. 

SC 42's international standards are essential for overcoming barriers to AI adoption while addressing ethical, societal, and technical concerns. The committee's ecosystem approach, collaboration with various organizations and focus on comprehensive standards contribute to the responsible and effective implementation of AI technologies across diverse sectors. The broad range of stakeholders who contribute to the work of SC 42 make ISO/IEC AI standards stronger because they take diverse perspectives into account. It also makes for a more comprehensive discussion about the challenge in front of us, which is to ensure that we have a broad and responsible adoption of AI. 

Free online workshops organized by SC 42 provide a platform for dialogue and information exchange on emerging trends, technology and international standards in the AI ecosystem. Find out more .