OpenAI is testing a feature that will allow ChatGPT to remember conversations. This information can be reused or retrieved in future conversations, which OpenAI says makes for more meaningful conversations.
ChatGPT currently sees every conversation as the first time it meets you. That’s because the generative AI chatbot has no memory. OpenAI is now launching an experiment to change that.
Optional and specific
Users are given several options for the feature. For example, it is possible to ask ChatGPT to remember only specific information. That could be about a part of the prompt that you keep repeating, such as placing titles in bold or your role and the company you work for. If a user chooses to keep track of all conversations, ChatGPT will also remember these details after several conversations.
It is always possible to ask ChatGPT what information it collected about you. Parts of the data can then be deleted. If you already know prior to a conversation that the information should not be kept, you can also change the conversation to a ‘temporary chat’. This option appears under the menu where you choose which GPT model to use. In the settings, users can turn off memory completely.
Source: OpenAI
Data remains private with Team and Enterprise
For users with a Team or Enterprise subscription, OpenAI assures that shared information remains out of ChatGPT’s training data. These users will get the same options to choose which information goes into memory and when memory is not used. Enterprise account administrators will also have the option to disable memory for the entire company.
Users of both subscriptions will still have to wait until memory is rolled out generally. “We are rolling out to a small portion of ChatGPT free and Plus users this week to learn how useful it is. We will share plans for broader roll out soon,” OpenAI adds.
More personal and time-efficient
The feature should make future conversations more meaningful and also save time when used in a business environment. For example, memory in ChatGPT ensures that the AI assistant always produces texts that take into account the company’s tone of voice. It further standardizes the output it generates based on previous prompts.
Also read: ChatGPT gets smarter, cheaper and less lazy