Overview
ISO/IEC TR 21221:2025 - Information technology - Artificial intelligence - Beneficial AI systems provides a conceptual framework for articulating and evaluating the benefits delivered by AI systems. The Technical Report focuses on how AI can generate functional, economic, environmental, social, societal, cultural, intellectual and personal value as perceived by different stakeholders. It also includes illustrative use cases and guidance for integrating benefit assessment with trustworthiness, purposefulness and sustainability (including links to the UN Sustainable Development Goals).
Key topics and technical focus
- Scope and definitions: Clear definitions for terms such as AI system, AI customer, AI user, value proposition, relevant interested party and sustainability.
- Conceptual framework: Attributes that make an AI system “beneficial” - functioning as designed, trustworthy, purposefully and responsibly applied, and able to create sustainable net positive value.
- Stakeholder perspectives: Roles and viewpoints of make (developers, producers, data providers), use (customers, users) and impact (community, regulators, civil society) stakeholders; models for mapping expectations and impacts (see Annex C).
- Trustworthiness considerations: Factors affecting stakeholder trust (explainability, robustness, transparency, ethics, data quality) and their influence on perceived benefits.
- Measuring AI value: Benefit components, automation effects, timeliness of benefits evaluation and cost‑benefit analysis approaches referenced for AI customers (Annex B).
- Positive impact and sustainability: Guidance on designing AI value propositions that support sustainable, net positive outcomes and contribute to the UN SDGs (Annex D).
- Use cases and modelling: Informative examples of beneficial AI system use cases (Annex A) and methods to model customer value.
Practical applications - who should use this standard
- AI developers and providers - to align system design with stakeholder value, trustworthiness and responsible-use criteria.
- Product managers and AI application providers - to craft compelling value propositions and business models that demonstrate measurable benefits.
- Procurement teams and AI customers - to perform cost‑benefit analyses and assess timeliness and sustainability of AI investments.
- Policy makers, regulators and civil society - to evaluate societal impacts, governance needs and contribution to public goals (e.g., UN SDGs).
- Quality, risk and ethics officers - to integrate benefit assessment with trust, explainability and lifecycle risk management.
Related standards
- ISO/IEC 22989:2022 - AI concepts and terminology
- ISO/IEC 5339:2024 - Guidance for AI applications
- ISO/IEC TR 24028 - Trustworthiness-related considerations (referenced for vulnerability and risk factors)
ISO/IEC TR 21221:2025 is a practical resource for organizations seeking to demonstrate and measure the real-world value of AI systems while embedding trustworthiness, responsibility and sustainability into AI strategy and governance. Keywords: ISO/IEC TR 21221:2025, beneficial AI systems, AI trustworthiness, AI stakeholders, AI value proposition, AI sustainability, UN SDGs, AI use cases.