AI 伦理
定义
AI 伦理涵盖原则（公平性、透明度、问责制、隐私）和治理 for 设计ing and deploying AI. It includes codes of conduct, impact assessments, and regulation.
It connects to AI safety (risks, alignment), bias in AI (fairness), and explainable AI (transparency). Regulation (例如 EU AI Act) is making ethics and impact assessments mandatory for high-risk applications. Organizations and practitioners need to operationalize principles into 设计, evaluation, and deployment practices.
工作原理
Organizations 采用指导方针 (例如 负责任的 AI 原则) and review processes (例如 ethics boards, impact assessments). Regulators set requirements (例如 EU AI Act: risk tiers, transparency, human oversight). Practitioners use checklists (例如 data provenance, fairness metrics), audits (例如 bias, safety), and stakeholder input to align systems with ethical norms. Flow: identify use case and risk level → assess impact (who is affected, what could go wrong) → implement safeguards (data, model, explainability, human-in-the-loop) → monitor and iterate. Documentation and accountability (who is responsible for what) are part of governance.
应用场景
当需要将系统与规范、法规和问责制对齐时，AI 伦理和治理适用（影响、公平性、隐私）。
- Impact assessments and governance for new AI products
- Aligning with regulation (例如 EU AI Act) and sector codes
- Privacy, fairness, and accountability in 设计 and deployment