数字企业责任边界扩展：从代码到价值链的全面尽责

南开大学人权研究中心与中国信通院联合发布了《数字企业及其供应链尽责管理指南》，系统阐述数字经济时代企业责任的新内涵1。该指南的核心创新在于将产品、技术和服务本身纳入人权尽责管理范畴，标志着企业责任理念从传统的行为责任向影响责任转变，从事后救济向事前预防转变1。

国际和国内监管环境对数字企业提出了更严格的标准。2024年6月，欧盟通过《企业可持续尽责指令》（CSDDD），对价值链进行了广泛界定，要求企业覆盖原材料采购、生产、运输到产品服务开发的全流程1。同时欧盟《人工智能法案》要求高风险AI系统建立风险管理体系、落实数据治理、设置人工监督机制1。在中国，新修订的《网络安全法》于2026年正式实施，首次在法律层面规定了人工智能专条1。

现实中已出现因数字系统管理不当引发的风险事件。2026年，Meta在美国遭集体诉讼，员工指控公司使用AI筛选系统导致休医疗假、申请残障便利的员工被不成比例地纳入裁员名单1。针对这些挑战，企业需做好三件事：责任到人把合规前置、将数字产品纳入风险管理、建立与本土供应链生态相衔接的评价能力1。该指南提出了覆盖企业运营、全球价值链和数字产品三个维度的尽责体系1。

Nankai University's Center for Human Rights Research and the China Academy of Information and Communications Technology jointly released the "Guidelines for Due Diligence Management of Digital Enterprises and Their Supply Chains" in July 2024, establishing a comprehensive framework for corporate responsibility in the digital economy era 1. The guidelines fundamentally expand how enterprises should understand and manage their obligations, integrating products, technologies, and services themselves into human rights due diligence processes 1.

The framework requires enterprises to transition from traditional conduct-based accountability toward impact-focused responsibility, shifting from reactive remediation to proactive prevention 1. The due diligence system encompasses three dimensions: enterprise operations, global value chains, and digital products 1. This approach aligns with regulatory developments globally, particularly the European Union's Corporate Sustainability Due Diligence Directive passed in June 2024, which mandates enterprises cover the entire value chain from raw material procurement through production, transportation, to product and service development 1. The EU's Artificial Intelligence Act similarly requires high-risk AI systems to establish risk management frameworks, implement data governance, and establish human oversight mechanisms 1.

Recent cases underscore the urgency of these standards. In July 2026, Meta faced a class action lawsuit in the United States, with employees alleging that the company's AI screening systems disproportionately included workers taking medical leave or requesting disability accommodations in layoff decisions 1. China's revised Cybersecurity Law took effect in 2026, introducing dedicated legal provisions on artificial intelligence for the first time 1. To address these challenges, experts propose that enterprises must prioritize three actions: embedding compliance through clear accountability structures, incorporating digital products into risk management frameworks, and developing evaluation capabilities that integrate with local supply chain ecosystems 1.