[隐私政策] How to use data to improve the performance of large models

OpenAI uses your data to improve and train its models. One of the most useful and promising features of AI models is that they can improve over time. Models are continuously improved through research breakthroughs and exposure to real-world problems and data. When you share your content with OpenAI, it helps OpenAI’s models become more accurate and better able to solve your specific problems, and also helps improve their general capabilities and safety. OpenAI will not use your content to market its services or create advertising profiles about you—it uses it to make its models more helpful. For example, ChatGPT improves by further training on conversations with the people who talk to it, unless you opt out.

For consumer services, such as ChatGPT and DALL•E
When you use consumer services such as ChatGPT or DALL•E, your content may be used to train OpenAI’s models.

You can opt out of training through OpenAI’s privacy portal by clicking “Do not train on my content,” or by turning off training for ChatGPT conversations according to the instructions in the Data Controls FAQ. Once you opt out, new conversations will not be used to train OpenAI’s models.

For enterprise services, such as ChatGPT Team, ChatGPT Enterprise, and the API platform
OpenAI does not use content from its enterprise services, such as ChatGPT Team, ChatGPT Enterprise, and the API platform, to train its models. For information about how enterprise data is handled, please see OpenAI’s Enterprise Privacy page.

What does the entire process look like?
OpenAI retains certain data from your interactions, but before using it to improve and train its models, it takes measures to reduce the amount of personal information in the training dataset. This data helps OpenAI better understand user needs and preferences, making its models more efficient over time.