📄 中文摘要
每个主要的 LLM 提供商（如 OpenAI、Anthropic、Google）都会收集、存储并利用你的 API 调用来改进模型。你的专有业务逻辑、客户对话、内部研究和产品策略都在他们的服务条款下成为了可利用的对象。这并不是一个漏洞，而是他们的商业模式。TIAMAT 隐私代理可以帮助你重新获得控制权。OpenAI 在未选择退出的情况下会将 API 调用用于训练数据，且即使选择退出，数据也会至少存储 30 天。Anthropic 的 Claude API 会存储对话数据以进行改进和安全审计。Google 的 Gemini API 则记录所有请求以进行分析、滥用防范和模型训练。
📄 English Summary
Why Your LLM Provider Wants Your Proprietary Data — And How to Stop Them
Every major LLM provider, including OpenAI, Anthropic, and Google, collects, stores, and utilizes your API calls for model improvement. Proprietary business logic, customer conversations, internal research, and product strategies are all fair game under their terms of service. This is not a bug; it is their business model. The TIAMAT Privacy Proxy offers a way to regain control. OpenAI uses API calls for training data unless you opt-out, and even then, data is stored for a minimum of 30 days. Anthropic's Claude API retains conversation data for improvement and safety auditing. Google's Gemini API logs all requests for analytics, abuse prevention, and model training.