At the recent biannual plenary for the joint IEC and ISO committee on AI, SC 42, held in Vienna, Austria, over 250 global experts from 60 countries convened to announce and discuss new initiatives in the field of artificial intelligence (AI). Notably, progress was made on key international standards, including the upcoming ISO/IEC 42001, set to be the world’s first certifiable management system standard for AI.
The event addressed future AI regulations, such as the impending European Union AI Act, emphasizing the role of international standards in ensuring compliance. Several new projects and collaborations were unveiled, including the establishment of two joint working groups focusing on functional safety in AI systems and natural language processing systems.
Additionally, the committee agreed on new study areas, encompassing topics like operational design domain, human-machine teaming, evaluation metrics for AI use cases, guidance on model training efficiency optimization, and an AI maturity model. Amendments were approved for foundational AI terminology and framework standards (ISO/IEC 22989 and ISO/IEC 23053) to incorporate generative AI concepts.
The creation of a handbook to aid small and medium-sized enterprises (SMEs) in utilizing the forthcoming ISO/IEC 42001 standard was also approved. Wael William Diab, Chair of the joint IEC and ISO committee on AI, highlighted the expansion of the program’s work as reflective of the increasing importance of international standards for responsible AI adoption.
The committee’s comprehensive approach considers both technical and non-technical aspects of AI, including business, regulatory, and ethical considerations. The strategic planning exercise undertaken by SC 42’s working groups focused on emerging areas for AI standardization, showcasing the committee’s commitment to staying at the forefront of AI developments.
Looking ahead, the fourth biannual ISO/IEC AI Workshop scheduled for December will cover four content tracks: AI applications, beneficial AI, novel AI standardization approaches, and emerging AI technology trends and requirements.