The biggest update to ChatGPT's memory system is now rolling out to users. OpenAI is expanding how the AI remembers information across conversations - moving from a system where users had to manually save specific facts to one that can draw on a much broader history of past interactions.
Until now, ChatGPT's memory worked like a shared notepad: you'd explicitly ask it to remember something ("remember that I prefer short answers" or "save that my business is a bakery"), and it would store those notes for future sessions. The upgrade lets ChatGPT pull context from your conversation history more broadly, without requiring you to flag anything manually. This is different from the context window - the amount of text an AI can process within a single conversation - because this is about the AI knowing you over time, across many separate sessions.
What This Means for Regular Users
For anyone using ChatGPT regularly for ongoing projects, this closes a real gap. If you've had a dozen conversations about building a client proposal, debugging a script, or developing a marketing campaign, the AI can now piece together context that previously lived in separate, disconnected sessions. The rollout is gradual - not all users have it yet, and paid plans appear to be getting access first.
The Privacy Trade-Off
More memory means more context, but it also means OpenAI's systems are processing a larger slice of your conversation history. Users who prefer to keep their sessions compartmentalized should check whether existing memory controls cover this new layer. ChatGPT does offer memory management settings, but their scope relative to this upgrade isn't fully clear yet.
For heavy users running recurring workflows, this is a meaningful change. The constant re-explaining - "here's the project brief, again, for the fifth time" - is one of the most consistent complaints about working with AI tools on multi-session tasks. If the upgrade works as described, that particular friction largely goes away.