Initiative overview
This standard was prepared by Joint Technical Committee ISO/IEC JTC 1, Subcommittee SC 42, Artificial intelligence. It addresses the fact that AI is increasingly applied across sectors using information technology and expected to be a main economic driver, with certain applications giving rise to societal challenges. Specific considerations include AI used for automatic decision-making, sometimes in a non-transparent and non-explainable way, requiring specific management beyond classical IT systems; the use of data analysis and machine learning rather than human-coded logic; and AI systems performing continuous learning that change behaviour during use, requiring ongoing attention to ensure responsible use continues.
The initiative is intended to be integrated with an organisation's processes and overall management structure. Crucial management processes include determination of organisational objectives and involvement of interested parties; management of risks and opportunities; management of trustworthiness concerns such as security, safety, fairness, transparency and data quality throughout the AI system life cycle; and management of suppliers, partners and third parties providing or developing AI systems. Organisations can apply a risk-based approach to ensure the appropriate level of control for particular AI use cases, services or products within scope, with organisational needs, size and structure, and the expectations of interested parties all influencing implementation.
It applies the harmonised structure developed to enhance alignment among management system standards, facilitating consistency with other standards related to quality, safety, security and privacy. It avoids specific guidance on management processes, instead allowing organisations to combine generally accepted frameworks, other International Standards and their own experience for processes such as risk management and data quality management. An organisation conforming with the requirements can generate evidence of its responsibility and accountability regarding its role with respect to AI systems.
Stakeholders in development include national bodies that are members of ISO or IEC, participating through technical committees, along with other international organisations, governmental and non-governmental, in liaison with ISO and IEC.
