Memory summaries and sources now visible

Improved memory for ChatGPT Business is now rolling out. When memory is enabled, ChatGPT can use relevant context from past chats to keep memory current and make responses more relevant as work changes, rather than relying only on details saved manually.

Users can now:

• Review a memory summary showing information ChatGPT may use to personalize responses
• View Sources below personalized responses to see relevant context from memories, past chats, and custom instructions
• Correct memory, delete a referenced chat, mark a source as not relevant, turn memory off, or return to legacy Saved memories

The memory summary is a high-level view of relevant context and may not include everything ChatGPT can remember or reference. This update does not affect Codex memory. Project-only memory remains contained within each project and does not use memories or conversations from outside that project. These improvements are available at no additional cost.