OpenAI is giving ChatGPT a better memory.
The San Francisco artificial intelligence start-up said on Tuesday that it was releasing a new version of its chatbot that would remember what users say so that it could use that information in future chats.
If a user mentions a daughter, Lina, who is about to turn 5, likes the color pink and enjoys jellyfish, for example, ChatGPT can store this information and retrieve it as needed. When the same user asks the bot to “create a birthday card for my daughter,” it might generate a card with pink jellyfish that reads “Happy 5th Birthday, Lina!”
With this new technology, OpenAI continues to transform ChatGPT into an automated digital assistant that can compete with existing services like Apple’s Siri or Amazon’s Alexa. Last year, the company allowed users to add instructions and personal preferences, such as details about their jobs or the size of their families, that the chatbot should consider during each conversation. Now, ChatGPT can draw on a much wider and more detailed array of information.
“We think that the most useful assistants are those that evolve with you — and keep up with you,” said Joanne Jang, an OpenAI product lead who helps oversee its memory project.
Although ChatGPT can now remember previous conversations, it can still make mistakes — just like humans can. When a user asks ChatGPT to make Lina a birthday card, the chatbot might create one with a subtle typo such as “Haippy 5th Birthday! Lina!”
The company is first providing the new technology to a limited number of users. It will be available to people using the free version of ChatGPT as well as those who subscribe to ChatGPT Plus, a more advanced service that costs $20 a month.
OpenAI is also introducing to users on Tuesday what it calls temporary chats, during which conversations and memories are not stored.
ChatGPT has for some time offered a limited form of memory. When users chatted with the bot, its responses drew on what they said earlier in the same conversion. Now, the bot can draw on information from previous conversations.
(The New York Times sued OpenAI and its partner, Microsoft, in December, for copyright infringement of news content related to A.I. systems.)
The bot builds this memory by automatically identifying and storing information that could be useful in the future. “We rely on the model to decide what may or may not be pertinent,” said OpenAI research scientist Liam Fedus, referring to the A.I. technology that underpins ChatGPT.
Users can tell the bot to remember something specific from their conversation, ask what has already been stored in its memory, tell the chatbot to forget certain information or turn off memory entirely.
By default, OpenAI has been recording entire ChatGPT conversations and using them to train future versions of the chatbot. OpenAI said that it removed personally identifiable information from conversations used to train its technology. And users can choose to remove their conversations from OpenAI’s training data entirely.
But creating and storing a separate list of personal memories that can be brought up by the chatbot in conversations could raise privacy concerns. The company argued that what it was doing was not that much different from the way search engines and browsers stored the internet history of their users.