Now ChatGPT Knows What You Told It Last Summer

OpenAI has rolled out an update to ChatGPT that allows it to reference all past conversations alongside existing saved memories. The update is meant to improve the AI’s ability to deliver context-aware responses and is now available to all Plus and Pro users globally, except in the EEA, UK, Switzerland, Norway, Iceland, and Liechtenstein.

“We have greatly improved memory in chatgpt- it can now reference all your past conversations! This is a surprisingly great feature, imo, and it points at something we are excited about: AI systems that get to know you over your life and become extremely useful and personalised,” said OpenAI chief Sam Altman in a post on X. 

The updated memory feature allows ChatGPT to draw on users’ full interaction history to improve response relevance, including for writing support, learning assistance, and advice. “New conversations naturally build upon what it already knows about you,” OpenAI said in a post on X. “This makes interactions feel smoother.”

However, users retain full control over memory features. According to OpenAI, anyone who has previously opted out of memory is also automatically opted out of chat history referencing. “You can opt out of referencing past chats, or memory altogether, at any time in settings,” the company confirmed.

“Memory isn’t just another product feature. It signals a shift from episodic interactions (think a call center) to evolving ones (more like a colleague or friend),” said Noam Brown, research scientist at OpenAI. He added that there is still a lot of research to do, but it’s a step toward fundamentally changing how we interact with LLMs.

“Will GPT think worse of me based on that noob bash question I asked 7 months ago,” joked OpenAI co-founder Andrej Karpatrhy.

“I totally get why AI long-term memory is useful and, based on my testing, think many people will love it… But I actually don’t want the LLMs I use for work to chime in with personal details or subtly change their answers as a result of my past interactions. Boundaries are good,” Ethan Mollick, professor at The Wharton School, posted on X. 

The feature is expected to roll out to Team, Enterprise, and Edu users in the coming weeks.

OpenAI has notified users that they’ll see a specific message in the ChatGPT interface once the enhanced memory is available to them.

However, OpenAI is not the first to roll out such a feature. In February, Google introduced a similar capability where Gemini can recall past chats to provide more helpful responses. Whether you’re asking about something previously discussed or requesting a summary of an earlier conversation, Gemini uses relevant chat history to generate its reply.