OpenAI has announced the launch of a new memory feature in the ChatGPT robot, allowing the robot to store and remember specific information about users and use it to provide answers.
The company has started offering the new feature to Plus plan subscribers worldwide, excluding the European Union and South Korea.
The company stated in its tweet on its official account on the X platform that it will provide the “memory” feature to team plan subscribers, company clients, and custom GPT robots in the future, where they can access this feature from the GPT store and add it to the robot.
The new memory feature in ChatGPT works in two ways; the first is by providing it with specific information directly, such as the job, workplace, favorite programming language, children’s names, bosses’ names, etc. The second method allows the robot to record some details and information in the context of conversations with it.
The feature was first announced by the company in the middle of last February and was undergoing testing before being launched.
Users can add information they want ChatGPT to remember permanently to ensure alignment with answers at all times. Users have the right to edit or delete those “memories” or disable this feature completely from the settings if they wish, with the company implementing this feature to automatically disregard sensitive information in memory.
Although this feature may raise some privacy concerns for some users, it is supposed to make the ChatGPT user experience more personalized and reduce the rate of incorrect answers.
