OpenAI has rebuilt ChatGPT's memory to automatically extract facts about you from your chat history — no manual input required.
The new system runs in the background, synthesizing details from past conversations and updating a running profile of your preferences, habits, and context. It's a departure from the earlier memory feature, which required users to either explicitly tell ChatGPT to remember something or manage a list of saved facts by hand. The rebuilt version infers on its own and refines what it knows over time. The rollout begins with Plus and Pro subscribers in the United States.
The gap between opt-in memory and always-on inference is larger than it sounds. Users who skip the settings menu may not realize ChatGPT is quietly assembling a behavioral profile across every session. OpenAI isn't alone — Google's Gemini and others have been pushing persistent personalization features — but the passive, automatic approach raises the stakes on what it means to simply use a chatbot.
Whether this makes ChatGPT feel like a more capable assistant or a more attentive data collector depends almost entirely on how much you trust OpenAI and how recently you've visited the privacy settings.