> **来源：[研报客](https://pc.yanbaoke.cn)** # OECD Due Diligence Guidance for Responsible AI Summary ## Core Content The **OECD Due Diligence Guidance for Responsible AI** is a comprehensive framework designed to help enterprises implement the OECD Guidelines for Multinational Enterprises (MNE Guidelines) and the OECD Recommendation on Artificial Intelligence (AI Principles). It outlines a structured approach to identifying, addressing, and mitigating risks associated with the development and use of AI systems across the entire value chain. This guidance is intended for **multinational enterprises** that are involved in the AI system value chain, including: - **Suppliers of AI inputs** (e.g., data providers, compute providers, cloud services) - **Developers of AI systems** (e.g., those involved in planning, design, model building, deployment, and operation) - **Users of AI systems** (e.g., financial institutions, real economy enterprises) The framework is based on the **OECD RBC due diligence approach**, which includes six key steps: 1. **Embed RBC into policies and management systems** 2. **Identify and assess actual and potential adverse impacts** 3. **Cease, prevent, and mitigate adverse impacts** 4. **Track implementation and results of due diligence activities** 5. **Communicate actions to address impacts** 6. **Provide for or cooperate in remediation when appropriate** The guidance emphasizes a **risk-based approach** to responsible AI, ensuring that enterprises take proactive steps to prevent harm while fostering innovation and trust. It also promotes **policy coherence** and **interoperability** between the MNE Guidelines, AI Principles, and other national or international AI risk management frameworks. ## Main Viewpoints - **Responsible AI is not just a technical issue**, but a broader ethical and social responsibility that involves all stakeholders. - **AI systems can significantly benefit society**, but they also pose risks that must be proactively managed. - **Stakeholder engagement and transparency** are central to the due diligence process, especially for workers and affected communities. - **The MNE Guidelines and AI Principles are complementary**, and this guidance bridges the gap between them by offering practical implementation steps. - **SMEs face unique challenges** in implementing due diligence due to limited resources and capacity, and the guidance encourages collaborative approaches and industry initiatives to support them. ## Key Information ### Scope of the Guidance - **Targeted at enterprises** across the AI system value chain, from suppliers to users. - **Not limited to the technology sector**, as AI is increasingly used in non-technological contexts. - **Excludes hardware supply chains** (e.g., mining and manufacturing of components) which are covered in separate guidance. ### Structure and Implementation - **Six steps** guide the due diligence process, each with practical examples and a roadmap of related provisions from existing frameworks. - The guidance is **not a checklist**, but a **tool for adaptation**, recognizing that different situations may require different approaches. - **Practical examples** are drawn from various AI risk management frameworks and industry consultations. ### Special Considerations - **SMEs** are encouraged to use **collaborative strategies** and engage in **industry initiatives** to reduce costs and improve compliance. - **Stakeholder engagement** is emphasized, particularly with workers and affected communities. - **Transparency, explainability, and traceability** are critical in the AI system lifecycle. - **Remediation** and **cooperation** are important when adverse impacts occur or are suspected. - **Interoperability** and **global cooperation** are key to aligning with international standards. ### Legal and Regulatory Context - The guidance draws on **existing international and national AI frameworks**, regulations, and initiatives. - It supports **National Contact Points (NCPs)** in promoting responsible AI standards and addressing violations. - It is also relevant for **civil society, workers, trade unions, and regulatory authorities** involved in AI oversight. ### Strategic Importance - Responsible AI is **increasingly a business requirement**, as more enterprises integrate AI risk management into their procurement and operations. - Demonstrating commitment to responsible AI can lead to **competitive advantage**, **access to capital markets**, and **trust among investors and regulators**. - The guidance helps **navigate complex regulatory environments**, ensuring that enterprises can operate effectively across multiple jurisdictions. ## Conclusion The OECD Due Diligence Guidance for Responsible AI is a **practical and adaptable tool** for enterprises seeking to align with international standards on AI ethics and governance. It supports a **holistic, risk-based approach** to AI development and use, promoting **trust, innovation, and sustainable growth** while addressing **ethical, social, and economic risks**. By following this guidance, enterprises can better meet the expectations of stakeholders, regulators, and investors, ensuring that AI technologies contribute positively to society.