📄 中文摘要
每次调用 <code>chat.completions.create</code> 时，都会将提示发送到 OpenAI 的服务器。如果提示中包含用户数据，例如支持票、表单输入或 CRM 记录，则很可能会包含姓名、电子邮件、电话号码等敏感信息。为了保护用户隐私，开发者需要采取措施，确保在与 OpenAI API 交互时不泄露个人身份信息。可以通过数据脱敏、过滤敏感信息等方式来降低风险，确保用户数据的安全性和隐私性。开发者应当重视这一问题，采取适当的防护措施。
停止通过 OpenAI API 调用泄露个人身份信息
出处: Stop Leaking PII Through Your OpenAI API Calls
发布: 2026年3月5日
📄 English Summary
Stop Leaking PII Through Your OpenAI API Calls
Every call to <code>chat.completions.create</code> sends the prompt to OpenAI's servers. If the prompt contains user data, such as support tickets, form inputs, or CRM records, it likely includes sensitive information like names, emails, and phone numbers. To protect user privacy, developers need to implement measures to prevent the leakage of personally identifiable information (PII) when interacting with the OpenAI API. Strategies such as data anonymization and filtering sensitive information can help mitigate risks and ensure the security and privacy of user data. Developers should prioritize this issue and adopt appropriate protective measures.