OpenAI is incorporating “memory” feature for ChatGPT.

Is it a “needed” by us or just another company trying to monetise by learning our “personality”

OpenAI announced that they are testing memory incorporation with ChatGPT. OpenAI wants ChatGPT to remember the discussions across different conversation we had with it. And based on the memory gained, ChatGPT will make the future conversation seamless.

It is true that, remembering things you discuss across all chats saves you from having to repeat information and makes future conversations more meaningful.

OpenAI, in its official blog, states the user is still in control of this memory.

We can either write scripts with details about us for the ChatGPT to “remember” or one can have settings to make ChatGPT to automatically take some interaction details to build its “memory”.

We can also ask ChatGPT to “forget” via conversation or again through settings.

ChatGPT at present behaves like “Leonard” from Christopher Nolan’s Memento. Every time, it behaves like a new person. Interacting with it is like having to prompt with the same Polaroid pictures or worse it may feel like to take up tattoos. OpenAI claims this wont be the case with this “memory” rollout.

Will it Spook users ?

Benefits are there, if ChatGPT remembers about you and knowing what to present to you so that it can serve you better.

However, many still have Dejavu moment when a sneaker add comes in their social network/google, when they have mentioned about that in a different platform.

Some even have claimed that they did not “write” it in any social media. However, they talked with their friends and their phone was with them.

Whether the tech giant apps are silently listening to our conversation and suggesting adds is a topic which often surfaces with friends & families.

Some say, may be its not you but some one in your circle has “written” in the tech giants apps. Armed with that knowledge, the adds are targetted to wider circle which includes you. And these was happening years before the latest AI revolution.

In any case, humans tend to get spooked when adds land within their private conscious boundary, matching their intention.

Hence, I believe many may be getting spooked with some response got from a information you shared with a powerful AI couple of years back.

Whats OpenAI Offers?

OpenAI in its official blog gives an example of this memory feature as follows,

“You mention that you have a toddler and that she loves jellyfish. When you ask ChatGPT to help create her birthday card, it suggests a jellyfish wearing a party hat.”

The roll out will be done initially to small portion of ChatGPT free and Plus users. Later it will be for broader user group.

What “may” Spook you

As said before, if one mentioned about their toddler a year back and totally forgot about the OpenAI conversation, they will get spooked when they receive a birthday card with jellyfish character.

We all know human’s memory depreciate faster than an electric car value

We mistrusting AI, owing to our poor memory, for eavesdropping is one thing.

But we may actually believe that we “only” gave that information, even when AI actually eavesdropped or combined multiple details to fabricated a truth / lie.

The reality will get blurred if AI starts to fabricate. Considering multiple “hallucinations” reported for AI, there is a non-zero possibility for this fabrication to occur.

Hence, it will be kind of scary when a powerful AI entity remembers about our personality and later “profiles” us to share some spooky advice or service.

Its even scary to think whether ChatGPT will get trained using our “personality” from these “memories”

Conclusion

I know, there will be multiple questions in your mind now.

Two critical question can be topping that list. For the 1st one, The answer is “Yes — By Default the memory is turned ON “ For the 2nd one, I dont know whether the ChatGPT actually “forgets” if you ask it to forget. . .

Even if we ask ChatGPT to forget, may be the memory will be erased. However, erasing the training obtained from the memory is tricky

If you still have questions, you have a easy option,

We can ask ChatGPT :-)

Follow me for more stories on innovation & deep tech. Do you know you can give 50 claps for a story at a time, if you adore it ???

This story is published on Generative AI. Connect with us on LinkedIn and follow Zeniteq to stay in the loop with the latest AI stories. Let’s shape the future of AI together!