April 12, 2025 — OpenAI has officially begun rolling out a major upgrade to ChatGPT: long-term memory. This new capability allows the AI to remember facts, user preferences, and previous conversations across sessions — a significant leap toward more personalized and intelligent interactions.
The memory feature, which has been in testing phases for several months, is now gradually being enabled for ChatGPT users. According to OpenAI, it will be turned on by default for all Pro users by mid-2024, with broader availability expected in the future. The memory upgrade marks a shift from the traditional session-based model — where ChatGPT forgets everything once a conversation ends — to a more continuous, evolving interaction that adapts to the user’s needs over time.
How the Memory Works
With long-term memory, ChatGPT can retain useful information to make future chats smoother and more relevant. For example, if a user tells the AI their name, preferred tone of conversation, or their profession, ChatGPT will be able to recall that information in later chats. It might remember that a user is a teacher who prefers concise explanations or that they’re working on a novel and like weekly check-ins on progress.
This ability enables a more dynamic user experience where the AI doesn’t need to be reminded of key details in every new session. It’s particularly helpful for people who use ChatGPT regularly for personal projects, business tasks, or learning goals.
Memory Management: Transparency and Control
To maintain transparency and give users full control, OpenAI has implemented robust memory management options. Users can easily view what ChatGPT remembers about them by going into the settings and accessing the “Memory” section. From there, they can delete individual memories or wipe all stored information at any time.
OpenAI has also made it clear that users will be notified whenever a memory is created or updated. A message like “I’ve added this to your memory” will appear in the chat when a new piece of information is saved. This helps users stay informed and in control of what the AI retains.
For those who prefer not to use this feature, memory can be completely disabled in the settings. Even with memory turned off, users can continue to benefit from ChatGPT’s core capabilities, just without the long-term context.
Gradual Rollout and User Feedback
The memory feature is being introduced gradually, allowing OpenAI to gather feedback and fine-tune the experience. The company has emphasized that this cautious rollout is intentional to ensure a smooth user experience and to make sure privacy and safety concerns are addressed properly.
In the coming months, more users will receive access to memory, with Pro subscribers prioritized in the rollout. OpenAI believes this feature will significantly improve the usefulness of ChatGPT, especially for those who rely on it as a digital assistant, writing coach, or learning companion.
A Step Toward Smarter AI
OpenAI’s introduction of long-term memory is part of a broader push to make AI more adaptive and intuitive. As ChatGPT continues to evolve, the company envisions a future where the AI becomes a truly helpful assistant, capable of growing and learning alongside its user.
This upgrade also sets the stage for more advanced features in the future, such as memory-driven task management, personalized learning paths, and smarter project collaboration tools.
While concerns about privacy and data handling remain central to discussions about AI, OpenAI’s approach of user-controlled memory, transparency, and opt-out options signals a step in the right direction.
With long-term memory now a reality, ChatGPT is moving closer to becoming a truly intelligent and personalized AI assistant — one that remembers not just what you said, but who you are.
