📄 中文摘要
2026年2月19日，微软的媒体诚信报告、加利福尼亚的数字尊严法案（SB 1142）和明尼苏达的去性别化禁令（HF1606）在24小时内相继发布。这三项发展均涉及对AI生成内容的问责，但都存在一个共同的盲点，即未能解决“负证据问题”——即证明AI系统确实拒绝生成有害内容。该研究对每项发展进行了基于主要来源的事实核查，将技术差距映射到CAP-SRP（内容/创意AI档案 - 安全拒绝来源）规范，并提供了构建加密拒绝审计轨迹的Python代码。
核实 AI 安全差距：微软媒体诚信报告，加利福尼亚数字尊严法案
📄 English Summary
Fact-Checking the AI Safety Gap: Microsoft's Media Integrity Report, California's Digital Dignity Act
On February 19, 2026, three developments emerged within 24 hours: Microsoft's media integrity report, California's Digital Dignity Act (SB 1142), and Minnesota's nudification ban (HF1606). All three address accountability for AI-generated content but share a significant blind spot: none tackle the 'negative evidence problem'—the challenge of proving that an AI system actually refused to generate harmful content. This article fact-checks each development against primary sources, maps the technical gaps to the CAP-SRP (Content/Creative AI Profile – Safe Refusal Provenance) specification, and provides working Python code for building cryptographic refusal audit trails.