OpenAI recently unveiled a significant update to its flagship ChatGPT application, introducing a persistent memory feature. This new capability is gradually being made available to select users of the free version and the premium ChatGPT Plus service, priced at $20 monthly. With this enhancement, ChatGPT can now retain information shared by users over multiple conversations, a departure from its previous session-based memory limit.
Previously, ChatGPT would start each new session with no recall of past interactions, except for any predefined instructions by the user. The introduction of persistent memory aims to streamline user interactions by eliminating the need to reiterate information in subsequent chats.
Initially, this feature is being introduced to a limited group of free and Plus users to gauge its effectiveness. The capability will be extended to Teams and Enterprise subscribers at a later date, reflecting OpenAI’s commitment to staying competitive in the rapidly evolving AI assistant market.
This strategic enhancement follows Google’s recent efforts to highlight its AI chatbot capabilities, including a significant rebrand and the introduction of its advanced Gemini model under a new paid plan.
OpenAI’s blog post details how the memory function will gradually learn from user interactions, allowing for more personalized and efficient communication. Users can specifically request the chatbot to remember certain details, enhancing future conversations. This feature supports various applications, from personal preferences in article formatting to preferred programming languages and data visualization styles.
While the collected memories from free and Plus users may inform model training, OpenAI assures users of the option for temporary chats, which remain private and unutilized for model improvement.
Memory will be enabled by default, but users can opt-out or manage specific memories via the ChatGPT settings, ensuring user control over their data.
OpenAI emphasizes safety and unbiased memory retention, particularly avoiding unsolicited sensitive information storage. With plans for a wider rollout that includes Enterprise and Teams users, OpenAI is also exploring opportunities for businesses to integrate personalized memory functions into their GPT-based services, promising a more customized user experience.