OpenAI, the company behind ChatGPT, is introducing a new “memory” feature for the chatbot, allowing it to recall previous conversations. This means users won’t have to repeat information they’ve already shared.
In a blog post dated February 13, OpenAI revealed that this memory capability would initially be available to a select group of ChatGPT users, premium and free, later in the week. The company also mentioned that it plans to expand this feature to more users soon.
The memory feature is designed to enhance as ChatGPT interacts more with users. It will help the chatbot remember user preferences and details from past conversations, making its responses more personalized and relevant.
Users will have control over this feature, including options to tell ChatGPT to remember or forget certain information and even switch off the memory function entirely. Temporary chats, which won’t be saved in the chat history, and specific instructions for the chatbot to remember or forget details are also part of this update.
Screenshot from ChatGPT memory settings. Source: OpenAI
OpenAI mentioned that it might use the data shared with ChatGPT, including any remembered information, to continue training its language models. However, users can choose to turn off data collection.
The company is also focused on addressing potential biases with this feature and ensuring it doesn’t automatically remember sensitive information unless explicitly directed by the user.
OpenAI believes the memory feature will make ChatGPT more efficient by learning their work styles and preferences, especially for premium users. Developers will also get the option to create GPT builds with their unique memory settings.
Additionally, OpenAI launched an app store for ChatGPT pro subscribers in January, offering specialized AI chatbots for various tasks.
Meanwhile, OpenAI co-founder Andrej Karpathy announced his departure on February 14, stating on X that he left to work on personal projects and that no specific event influenced his decision.
