In the paper of “Scoping the OECD AI Principles” published in 2019 by OECD , there are five principles with AI system, including Inclusive and sustainable growth and well-being, Human values and fairness,Transparency, Robustness and safety, and Accountability. During the AI system development with the principle of human values and fairness,some bias can occur in different aspects, including data collection,technical bias, expert conception and environment. Risk management could be helpful to treat potential risks of adverse effects of AI system behavior. The procedures of risk management of AI system include objective setting, stakeholders and participators identification, riskevaluation, risk mitigation, implantation and monitoring.

Reference：https://bit.ly/3iOdWeZ