as training data. However, users are given the option to opt out by changing their settings (via Settings → Data Controls → Improve the model for everyone, and switching this option to Off). By contrast, for enterprise-oriented plans such as ChatGPT Team, ChatGPT Enterprise, and the API (developer use), the default setting is opt-out. In other words, user data is not used for training purposes under these plans. Google provides AI services not through a standalone Gemini pricing plan, but by integrating Gemini into other Google services such as Google Search, Google Workspace, and Google Cloud. Similar to ChatGPT, in the case of free and individual paid plans, data uploaded by users may, by default, be used for AI training purposes. For enterprise-oriented plans such as Google Workspace and Google Cloud, Google states that user data is not used as training data. Google also allows users to prevent their data from being used for model training by turning off the “Gemini App Activity” feature. In this case, conversations themselves are not stored. In other words, with ChatGPT, users can choose an option that prevents their data from being used for training without deleting conversation history, whereas with Gemini, opting out of training also results in the deletion of conversation records. In the case of Anthropic, the provider of the Claude service, a policy change introduced on October 8, 2025 allows users, at the time of sign-up, to choose whether their data may be used for AI training and improvement purposes. This setting can also be changed later through the user’s account settings. Anthropic likewise states that Generative AI Guide for Civil Society