While I worry about my privacy when using generative AI products like ChatGPT and Google Bard, I’d also like to be able to customize these chatbots so they offer a better personal assistant experience. That is, I’d love them to know things about me without sharing that information with the mothership, whether it’s OpenAI, Google, or someone else.


ChatGPT already made the first steps in that direction, introducing custom instructions in mid-July for Plus users. OpenAI then pushed the feature to free ChatGPT users in August. The feature lets you provide certain information to the chatbot that it’ll take into account for the prompts that fit. That way, you won’t have to provide the same information to chatbots in new prompts on similar topics.

Google is developing a Memory feature for Google Bard that will work similarly. You’ll be able to share some information about yourself with the chatbot. Bard will take it into account during prompts that are linked to that information.

I told you how I used ChatGPT to purchase running shoes and how I developed a half-marathon training plan using the same ChatGPT program. That was before the custom instructions arrived.

Now, I can just give the bot background information about myself. ChatGPT can use that information in any prompts dealing with exercise and nutrition. That way, I won’t have to repeat my age, weight, height, running pace, or medical conditions I have alongside every prompt that has to do with working out, eating, and health.

I’d also use ChatGPT’s memory feature to inform the bot that I don’t fully trust generative AI. Therefore, I want links and sources for every claim it makes in our conversations.

I think every ChatGPT alternative out there should offer similar features as long as these companies are mindful of user privacy. I’ll also point out that you’re in charge of what you share with ChatGPT, so don’t give it information that’s too personal or sensitive.

Google Bard might be the next ChatGPT rival that will remember what you like, thanks to the Memory features. Developer Dylan Roussel (via 9to5Google) discovered the feature.

Screenshots the developer obtained indicate that Google is working on the feature. The company offers a few examples of memories you’d like Bard to remember:

I try to avoid eating meat
I have 2 kids
Please give shorter responses

These could be important in a variety of cases. Some apply to specific prompts (avoiding meat as an ingredient for new recipes and meal plans). Others will apply to every Google Bard response (the shorter responses).

You’ll probably be able to add plenty of memories as you expand your Google Bard experience. And you’ll probably be able to remove memories that are not relevant again. Moreover, there’s a toggle that lets you turn memories on or off. This would be a great privacy feature, as you might not use your memories with all Google Bard prompts.

It’s unclear when the Google Bard Memory feature will roll out. As I pointed out before, new generative AI features roll out at an incredibly fast pace, so I wouldn’t be surprised to see Google’s chatbot get memories very soon.

[ For more curated tech news, check out the main news page here]

The post Google Bard’s Memory feature will let you teach the chatbot about yourself first appeared on

New reasons to get excited everyday.

Get the latest tech news delivered right in your mailbox

You may also like

Notify of
Inline Feedbacks
View all comments