OpenAI has announced a new update to ChatGPT that enables the AI to remember users’ previous conversations, aiming to enhance the model’s ability to deliver more personalized responses.
The new feature will appear in the app settings under the name “reference saved memories.”
Initially, this feature will be available to ChatGPT Pro and Plus subscribers, with the exception of users in the UK, the European Union, and four other European countries due to the need for additional legal reviews.
The main goal of this update is to improve the user experience by eliminating the need to repeat the same information in every new conversation.
According to OpenAI, the new memory capability adds deeper context to text and voice conversations, and even during image generation, which in turn enhances the interaction between the user and the model.
However, despite these benefits, the feature may raise concerns among users regarding the collection of personal information.
In response, OpenAI stated that the memory option can be disabled in the settings. Users will also be able to manage their saved memories or switch to temporary chat mode, which prevents any data from being stored.
Separately, Sam Altman – CEO of OpenAI – recently revealed future plans that include the launch of new models named “o3” and “o4-mini,” ahead of the highly anticipated GPT-5 release later this year.
He explained that these changes come as part of efforts to improve the architecture of GPT-5, which will enable the company to deliver a more advanced model.
Altman also noted that the upcoming “o3” version will offer significant improvements over previous test releases and is expected to receive positive feedback from users.
While he did not disclose specific details about the capabilities of the new models, Altman confirmed that the company remains focused on boosting AI efficiency and making it more adaptable to users’ needs.
