Related standards or drafts

• Artificial intelligence
• Machine learning
• (ai)
• Artificial intelligence — management system
• Artificial intelligence — overview of ethical and societal concerns
• Governance implications of the use of artificial intelligence by organizations
• Artificial intelligence — guidance on risk management
• Overview of trustworthiness in artificial intelligence
• Functional safety and ai systems
• Security and privacy in artificial intelligence
• Assessment of machine learning classification performance
• Guidance for quality evaluation of artificial intelligence (ai) systems

Artificial Intelligence (AI), a rapidly evolving frontier of the technology world, carries tremendous potential benefits but also notable risks, raising profound ethical, sociological and technological questions. Recognizing these challenges, an array of global organizations are formulating a plethora of international and European standards to streamline AI operations, ensuring safety, privacy, data protection and trustworthiness of AI-based systems.

Standards developed by the International Organization for Standardization (ISO), the International Electrotechnical Commission (IEC), the European Committee for Standardization (CEN) and the European Committee for Electrotechnical Standardization (CENELEC) provide significant standards, technical specifications and guidelines to foster consistency and safety in the burgeoning AI landscape.

Artificial Intelligence (AI) and Machine Learning (ML) are rapidly reshaping all aspects of our lives and industries. Ensuring safe, reliable, and ethical use of these technologies is critical, hence the emergence of numerous standards to address different aspects of their application. Some relevant standards include:

• CEN/CLC ISO/IEC/TR 24027:2023 exposes the issue of bias in AI systems, particularly those involved in decision-making processes. Bias assessment methods and measurement techniques are highlighted, covering all AI system lifecycle phases such as data collection, continual learning, design, training, and evaluation.
• CEN/CLC ISO/IEC/TR 24029-1:2023 provides insights on the robustness evaluation methods of neural networks in relation to AI. The document offers background details about the prevailing tactics designed for this purpose.
• CEN ISO/TR 22100-5:2022 delves into the safety implications of AI and ML for machinery and machinery systems, detailing the risk assessment process with respect to potential hazards associated with AI applications in such systems.
• EN ISO/IEC 22989:2023 establishes critical terminology and outlines key concepts in the AI field, benefitting standard development and facilitating communication among varied stakeholders.
• EN ISO/IEC 23053:2023 sets out an AI and ML framework, illustrating system components and their respective functions within the AI ecosystem. It is designed to be universally applicable to every type and size of organization that uses or implements AI systems.
• IEC SRD 63416:2023 ED1 describes the ethical considerations pertinent to the application of AI in the active assisted living (AAL) context, supplementing foundational AI ethical guidelines with AAL-specific provisions.
• ISO/IEC 23894:2023 offers guidance on managing risk specifically related to AI, for those who produce, use, or deploy products, systems, and services that harness AI.
• ISO/IEC 24668:2022 defines a process management framework to maximize big data analytics' benefits across organizations, addressing management processes for big data analytics and their interlinkages.
• ISO/IEC 38507:2022 introduces guidance for governing bodies to enable efficient, effective and responsible AI utilization within their organization, as well as addressing AI's implications for them.
• ISO/IEC 42001:2023 lays out requirements and counsel for formulating, implementing, maintaining, and improving an AI management system within an organizational framework.
• ISO/IEC 5338:2023 specifies the processes and related concepts for describing machine learning-based AI systems life cycle, assistive in AI systems development or acquisition.
• ISO/IEC 5339:2024 avails guidance for discerning the surroundings, opportunities, and methods for evolving and deploying AI applications from a broad perspective.

In conclusion, these standards encapsulate a comprehensive view of various aspects of AI and ML – from bias, safety, ethics, governance, lifecycle processes, risk management, and big data analytics. Each standard addresses different elements of AI and ML while collectively forming a coherent AI and ML framework for various entities ranging from businesses, governmental bodies, to not-for-profit organizations.
More information is provided on all CEN, CENELEC, ISO, IEC standards on Artificial Intelligence (AI) and Machine Learning, below.