Be a part of our each day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Study Extra
OpenAI is slowly rolling out higher reminiscence on ChatGPT, making it a default for ChatGPT to reference previous conversations. This has raised the worry that the platform is proactively “listening” to customers, making them uncomfortable with how a lot the platform is aware of.
ChatGPT already logs info from earlier interactions via its Reminiscence characteristic, making certain preferences are saved and conversations can seamlessly proceed from the place the person left off.
This new replace permits ChatGPT to “draw on past conversations to deliver more relevant and useful responses” and go throughout all modalities within the platform. Enhancements in Reminiscence enable future conversations, not simply present chat home windows, to reference earlier chats. It can solely be out there for ChatGPT Plus and Professional customers. ChatGPT Enterprise, Crew and Edu will get entry to the characteristic later.
Beginning at the moment, reminiscence in ChatGPT can now reference your whole previous chats to offer extra customized responses, drawing in your preferences and pursuits to make it much more useful for writing, getting recommendation, studying, and past. pic.twitter.com/s9BrWl94iY
— OpenAI (@OpenAI) April 10, 2025
OpenAI added Reminiscence to ChatGPT in February final 12 months to make speaking to ChatGPT extra useful. Reminiscence is a characteristic for many chat platforms and huge language fashions (LLMs). Gemini 2.0 Flash Considering added Reminiscence, whereas frameworks like A-Mem enhance long-context reminiscence for extra difficult duties.
Proactive reminiscence
Enhancements to reminiscence will let ChatGPT “naturally build” on earlier chat, and over time, OpenAI mentioned interactions on ChatGPT will probably be extra tailor-made to the person.
OpenAI provides two methods to manage Reminiscence via settings. The primary is Reference Saved Reminiscences, the place the person can direct ChatGPT to recollect info like names or preferences. The corporate mentioned folks normally add this info by explicitly telling ChatGPT to recollect one thing. The mannequin will determine which info will probably be useful in future conversations.
The second management is Reference Chat Historical past. This setting permits ChatGPT to attract context from earlier discussions and “adapt to your tone, goals, interests, or other recurring topics.” Nevertheless, the context won’t be saved or proven within the settings web page like saved recollections are.
“You can choose to have both settings on or off, or just turn on reference saved memories,” OpenAI mentioned. “The settings are flexible, and you can change them anytime, including managing specific saved memories. If you opt out, ChatGPT won’t draw on past conversations. You can also ask what it remembers or switch to Temporary Chat for memory‑free sessions.”
Considerations from some customers
Remembering conversations and taking particulars for future conversations not solely makes it simple to proceed a chat, however ideally, for enterprise duties, accessing preferences and context makes AI fashions extra helpful.
AI investor Allie Okay. Miller mentioned in a submit on X that this replace makes ChatGPT “listening all the time. It’s cutting across all of your conversations, whether you have explicitly asked it to remember something or not.”
“As I mentioned a few weeks ago, memory is the best feature inside these platforms. As models and features get commoditized, it’s going to come down to personalization, collaboration and network effects. Memory is the key. Memory is the moat,” Miller mentioned.
Let me clarify the brand-new OpenAI launch that stored Sam Altman up all evening final evening.
I haven’t got reminiscence turned on ChatGPT (as a result of I am unable to for work), and at the moment is the day that begins to harm.
Till at the moment, ChatGPT’s reminiscence was fairly boring. It waited for a transparent… https://t.co/fdAq1JLlwA
— Allie Okay. Miller (@alliekmiller) April 10, 2025
Nevertheless, after OpenAI introduced the Reminiscence replace, some customers expressed concern that it would change how the mannequin interacts with you.
Distinguished AI commenter and Wharton professor Ethan Mollick famous it’s not a characteristic he’ll activate.
“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 my LLMs I use for work to chime in with personal details or subtly change its answers as a result of my past interactions. Boundaries are good,” Mollick mentioned.
I completely get why AI long run reminiscence is beneficial and, based mostly on my testing, assume many individuals will adore it… however I really don’t need my LLMs I exploit for work to chime in with private particulars or subtly change its solutions because of my previous interactions.
Boundaries are good.
— Ethan Mollick (@emollick) April 10, 2025
OpenAI cofounder Andrej Karpathy fearful ChatGPT “think worse of me based on that noob bash question I asked 7 months ago.”
Will GPT assume worse of me based mostly on that noob bash query I requested 7 months in the past ?
— Andrej Karpathy (@karpathy) April 10, 2025
Reminiscence on ChatGPT will probably be useful, however will probably be as much as the person to find out how a lot they need the chat platform to find out about them and the way essential previous info will probably be for future conversations.
