On 10 April 2025, OpenAI unveiled a major update to ChatGPT that effectively gives the AI a memory – and marks the beginning of a new era in human-machine interaction. With the enhanced memory feature, ChatGPT can now recall previous conversations, analyse them, and use this context to deliver significantly more personalised and relevant responses. What was once just a chronological chat history now becomes a long-term foundation for more intelligent dialogue – dynamic, tailored, and continuously learning. The key upgrade: ChatGPT can draw on all previous interactions, not just explicitly saved memories but also automatically gathered chat patterns, provided the user has enabled this. The result is an AI that doesn’t just respond, it remembers.
This new memory function is more than a clever gimmick – it’s a strategic step toward a truly personal assistant. ChatGPT now remembers your preferred writing style, interests, and how much detail you like in your answers. If you regularly discuss technology, marketing or psychology, you’ll start receiving responses that go deeper, sound more like you, and align better with your knowledge level. Interactions become faster, more efficient, and more natural. At the same time, users remain firmly in control of their data. Those who don’t want ChatGPT to remember anything can turn the memory off, delete specific memories, or use the new “Temporary Chat” mode, which stores no data at all.
Still, this update raises important questions, especially in regions with strong data protection regulations. Since memory is processed on OpenAI’s servers, issues around privacy and personal data security are front and centre. As a result, the feature is currently not rolling out in the EU, UK, Switzerland, Norway, Iceland or Liechtenstein, where regulatory reviews are still ongoing. Elsewhere, the rollout is already underway, though initially limited to Plus and Pro subscribers. In the coming weeks, the memory function will also be made available to Team, Enterprise and EDU accounts.
OpenAI is also placing a strong emphasis on transparency and user choice. In the updated settings menu, users can see what the AI has remembered about them, edit it, or wipe it entirely. By default, conversations are still used to help improve the model, but users can opt out of this in the settings. It’s part of OpenAI’s attempt to strike a balance between meaningful personalisation and responsible data handling.
This update marks a new chapter for ChatGPT – one in which the AI evolves from a mere tool into a true personal assistant. It’s a glimpse of what many believe is the next logical step in AI development: systems that don’t just complete tasks, but think along with you, remember your needs, and adapt to you over time. Whether that vision becomes reality will depend not just on technology, but on trust. What’s clear already: the foundation has been laid – and the machine’s memory has been switched on.
