By Giving ChatGPT A Memory, OpenAI, Wants It To 'Remember Things You Discuss'

OpenAI Memory

Talking to an generative AI-powered chatbot can fun, amusing and astonishing. But after a while, things can be a burden.

This happens as users have to tell the AI their preferences, for more than several times. This can be groundhog, or annoying to say the least. A smart AI should be smart, and that it should remembers the things it's told, without having users to repeat themselves, over and over again.

OpenAI, the company known for its ChatGPT chatbot, knows this very well.

This is why the company gave the AI a 'Memory' feature to remember past conversations.

At the time, it was only available to a "small portion."

This time, Memory is made available to all ChatGPT Plus users, except those in Europe or Korea.

Memory works in one of two ways.

First, users can tell ChatGPT to remember something specific about them.

Like for example, a programmer who always write code in JavaScript, who work in a company is led by a man called "bro," and has an allergic to crabs.

Second, users can let ChatGPT to simply try to pick up those details over time, storing information about them as they ask questions and get answers.

In either case, the goal is for ChatGPT to feel a little more personal and a little smarter, without needing to be reminded every time.

Each custom GPT users use will have its own memory, too.

For example, OpenAI uses the Books GPT as an example.

When Memory is turned on, it can automatically remember which books the user has read, and which genre they like best.

Memory is a feature ChatGPT desperately needs.

But there are issues.

By design, Memory is like giving the AI a storage space to remember its user. But to do this, it needs to send users' conversation to its servers, and let its sophisticated system develop a profile of its user over time.

What this means, Memory is full of tidbits about users, their conversations, their preferences, and pretty much, all of their life.

This can make a lot of people, especially those who concern privacy, to feel really uncomfortable.

Many users are already wary of having their questions and missives hoovered up by OpenAI and fed back into the system as training data to help personalize the bot even further.

The the idea of ChatGPT "knowing" users is both awesome and creepy.

This is why OpenAI said that users are always in control of ChatGPT’s memory.

"You can explicitly tell it to remember something, ask it what it remembers, and tell it to forget conversationally or through settings. You can also turn it off entirely," the company said.

What's more, OpenAI said that its system is not designed to remember sensitive things like information about health.

OpenAI is also pitching Temporary Chat as a sort of incognito mode: a way to have a quick conversation without affecting ChatGPT’s memories of the user.

Users can also just turn it off entirely across their entire account.

Before introducing Memory, ChatGPT has what it calls the 'Custom Instructions', which allows the chatbot to get to know about users, and also prevent users from repeating themselves.

Published: 
30/04/2024