ISO/IEC TR 24027:2021 Information Technology - Artificial Intelligence (AI) - Bias in AI systems and AI aided decision making
- Publisher: ISO
- Status:
final - Version:
1 - Certifiable: False
- Release Date:
2021-11-01 - Date Added:
2025-09-23 - Source URL: https://www.iso.org/standard/77607.html
Summary
Provides technical guidance on identifying, assessing, and mitigating bias in AI systems and AI-assisted decision making. The report describes different types of bias and their potential impact on fairness, transparency, and trustworthiness. It is intended as a reference for practitioners and policymakers to incorporate bias considerations into design, development, evaluation, and deployment of AI systems.
Key Takeaways
- Defines key categories of bias, including data imbalance, representation bias, measurement bias, and algorithmic bias.
- Explains how bias can arise at different stages of the AI lifecycle — from data collection and preprocessing to model training, validation, and use.
- Emphasizes the link between bias and ethical principles such as fairness, accountability, and non-discrimination.
- Provides strategies and techniques to mitigate bias, such as improved data governance, algorithmic audits, and diverse stakeholder engagement.
Additional Sources
Tags
bias, risk-management, fairness, accountability, transparency