Anthropic announced that its AI chatbot Claude will add a new feature of "quoting past conversations", but it will not automatically record all interactions. Instead, it will adopt a more cautious "on demand" approach and will only call up past conversation content when the user actively requests it.
Anthropic explains that this feature can help users reopen a project after a period of time or find previously researched data without having to repeatedly re-explain the context. Claude will only retrieve relevant conversation records when explicitly requested, and the scope of reference is limited to the current scope and project.
In contrast, OpenAI updated itsChatGPT memory function, all conversations will be automatically saved and these records will be used to make personalized adjustments in future responses, so that each interaction can affect the subsequent answers.Similar functions, and has even tested using Google search history to further enhance personalized AI responses.
In contrast, Claude's approach is more like a keyword search for chat history rather than building a complete personal profile. Anthropic says this design is more privacy-oriented, and users can choose to completely turn this feature off in the settings to ensure that no past content is accidentally retrieved.
Currently, this feature is available to users of Claude Max, Team, and Enterprise subscription plans, and is expected to be expanded to more plans in the near future.
Anthropic emphasized that this change isn't intended to make Claude a long-term "user memory bank," but rather to provide a way to quickly retrieve previous conversations while balancing privacy and convenience. For businesses and professional users who value data control, this may help strike a balance between efficiency and security in AI applications.
