📄 中文摘要
欧盟人工智能法案第12条要求对每个高风险人工智能决策进行防篡改日志记录。2026年8月2日起，所有高风险人工智能系统的运营者必须遵守这一要求，特别是在金融、医疗、就业筛选、关键基础设施和公共服务等受监管领域。该条款具体规定了事件的自动记录，确保所有决策过程的透明性和可追溯性。合规不仅仅依赖于现有的监控工具，必须建立全面的日志记录机制，以满足法规要求。
欧盟人工智能法案第12条：AI代理日志记录的实际意义（附代码示例）
📄 English Summary
EU AI Act Article 12: What AI Agent Logging Actually Means (With Code Examples)
The EU AI Act Article 12 mandates tamper-evident logging for every high-risk AI decision. Starting from August 2, 2026, operators of high-risk AI systems, particularly in regulated sectors such as finance, healthcare, employment screening, critical infrastructure, and public services, must comply with this requirement. The article specifies the need for automatic logging of events to ensure transparency and traceability in decision-making processes. Compliance cannot rely solely on existing monitoring tools; a comprehensive logging mechanism must be established to meet regulatory demands.