Anthropic has introduced a new feature for its Claude AI chatbot that enables the AI to automatically remember details from past conversations for Team and Enterprise users. This capability allows Claude to integrate information regarding user preferences, project contexts, and prioritized tasks into its responses without requiring explicit prompts from users.
Previously, Anthropic offered paid users the ability to prompt Claude to recall previous discussions, but the latest update streamlines this process by allowing automatic retrieval of pertinent details. Claude’s memory function also extends to various projects, enabling users to create materials such as diagrams, website designs, and graphics based on uploaded files.
The memory feature appears to focus primarily on work-related aspects such as team processes and client needs. Anthropic emphasizes that the memory function is entirely optional for users, providing flexibility in how it can be utilized.
Users can manage Claude’s stored memories through a settings menu, where they can view and modify the information the AI retains. According to the company, modifications made by users on what Claude should focus on or disregard will directly influence the AI’s memory reference.
Other companies, including OpenAI and Google, have also implemented memory features in their chatbots. A recent article from The New York Times noted a correlation between the introduction of such memory capabilities and an increase in reports of misleading or “delusional” responses from chatbots like ChatGPT.
In addition to the memory functionality, Anthropic is launching incognito chat options for all users, allowing discussions that Claude will not save or reference in future conversations. This mirrors a similar feature—Private Chats—recently introduced by Google for its Gemini AI.
Source: https://www.theverge.com/news/776827/anthropic-claude-ai-memory-upgrade-team-enterprise