OpenAI has introduced a major upgrade to ChatGPT by enhancing its memory capabilities, allowing it to reference past conversations for more personalized and context-rich interactions. Previously, memory was limited to manual entries or short-term context within the same chat. Now, ChatGPT can recall insights from multiple sessions, enabling it to assist more intelligently over time. This feature is designed to make the AI feel more like an ongoing assistant rather than a one-off tool.
The memory update works in two ways: users can manually input information they want ChatGPT to remember, and the model can also autonomously collect and retain relevant details from past interactions. This allows it to remember names, preferences, ongoing projects, and more. For example, if you often ask it to write in a formal tone or reference a specific project, ChatGPT will adjust its responses accordingly without being reminded.
OpenAI is rolling this feature out gradually to Pro, Plus, Team, Enterprise, and Education subscribers, with free-tier users expected to gain access later. However, due to regional privacy regulations, the feature is not yet available in the EU, UK, and a few other countries. This phased rollout ensures user feedback and compliance can be incorporated effectively.
Importantly, users have full control over ChatGPT’s memory. You can view what the AI remembers, delete specific memories, or turn memory off entirely in the settings. Temporary chats also offer a privacy-first mode that doesn’t save any data. OpenAI emphasizes transparency and user agency, framing this feature as a key step toward building AI that evolves to support users over the long term.
