📄 中文摘要
针对 OpenAI 和 Anthropic 的诉讼将训练数据污染问题从一个小众的基准测试问题转变为生成性 AI 的核心法律和监管焦点。最初对基准测试膨胀的担忧，现在被框架为对个人和受保护数据在互联网规模上进行非法处理、保存和披露的指控。欧洲监管机构已将生成性模型视为2026年数据保护制度中最复杂的挑战之一，因为这些模型吸收了大量的个人和敏感数据。
关于针对 OpenAI 和 Anthropic 的 AI 训练数据污染诉讼的内部分析
📄 English Summary
Inside The Ai Training Data Contamination Lawsuits Targeting Openai And Anthropic
Lawsuits against OpenAI and Anthropic are transforming the issue of training data contamination from a niche benchmarking concern into a central legal and regulatory flashpoint for generative AI. What started as worries about inflated benchmarks is now framed as allegations of unlawful processing, retention, and disclosure of personal and protected data at internet scale. European regulators already view generative models as one of the most complex challenges for the 2026 data protection regime, given their absorption of vast quantities of personal and sensitive data.