Well, if you had ever wondered if ChatGPT remembers your conversations, here’s the answer – it does!

With a new memory upgrade, ChatGPT will be able to remember past conversations and recollect them as and when needed to provide meaningful outcomes. The creator of ChatGPT, OpenAI, claims that this upgrade will help the chatbot to know its users over a lifetime. Last year, OpenAI had added “Memory” feature to ChatGPT to retain limited information, such as prompts, queries and customizations, for future purposes. The new update differs in a couple of ways.

First, the Memory feature will allow ChatGPT to recall information that user manually ask ChatGPT to add to “saved memories”.

Second, the feature will enable ChatGPT to use “reference chat history” comprising of insights that ChatGPT gathers using past chats to refine the new conversations.

This new feature will be made available in all regions worldwide, except in the European Union, the United Kingdom, Norway, Switzerland and Liechtenstein. These countries have strict regulations in place that impede the implementation of such features in AI-powered chatbots. Only the customers paying for ChatGPT’s $200 monthly Pro subscription will be able to use this feature, which will later be rolled out to Plus subscribers paying $20 a month. The Teams, Enterprise and Edu users will get access to this feature soon, with no date specified so far.

OpenAI has clarified that the feature won’t be imposed over anyone. Users will have an option to toggle off saved memories under the personalization settings. However, it does raise some ethical concerns around the chatbot as users may not always like to get their conversations saved and may be unsure if ChatGPT is retaining information despite turning the toggle off. Moreover, if and how the saved memories are used by the company for improving ChatGPT platform is also unknown. Will the saved memories be private to the user, or will the company be able to peek into them? And what if hackers gain access to such private conversations between humans and chatbots? How big a breach of privacy will it be?

There are pertinent questions remaining unanswered. However, we must acknowledge that the AI-powered chatbots are growing increasingly capable of behaving like a human mind, and addition of long-term memory is certainly going to take the bar far higher for AI chatbots.