Unverified85% confidenceFactTime unknown
如果App使用了生成式AI或机器学习模型处理用户输入内容，开发者必须在隐私政策中披露数据是否会被发送至第三方服务器、是否用于模型训练、以及用户如何控制自己的数据
1
Sources
85%
Confidence
Long-term
Relevance
6/3/2026
First Seen
Sources
Related Entities
Related Claims
Unverified开发者应在合同层面要求 AI 提供商提供「数据不用于训练」的书面承诺，并在隐私政策中列出所有第三方 AI 服务商名称及数据处理地区73% similarUnverifiedThe free-tier terms of most AI service providers reserve the right to use user input data for model training, whereas enterprise editions typically offer a commitment that data will not be used for training.72% similarUnverifiedIn trade secret determinations in the AI field, model architecture may be protected by copyright, while training methods, data recipes, hyperparameter tuning strategies, and proprietary evaluation benchmarks are typically protected as trade secrets.72% similarUnverifiedSome AI tools use user interaction data for model fine-tuning training by default, requiring users to actively go into settings to disable this.69% similarUnverified端侧推理（On-Device Inference）在用户设备本地完成AI计算而不上传原始数据，可能成为平衡隐私与功能的折中方案，但对模型大小和设备算力要求更高66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/39050API
curl https://kongchang.com/api/v1/knowledge/claims/39050MCP
get_claim(id=39050)