Google is working on a new Bard feature that will allow the AI to remember its users’ preferences. Aptly called Memory, the latter allows you to give indications to the chatbot which it will have to remember and take into account when formulating its responses.
Bard, Google’s answer to ChatGPT, doesn’t know who you are. This may seem obvious – and even reassuring given Google’s annoying habit of collecting our personal data – but this detail is far from trivial when we try to chat with AI. Indeed, if you start a new conversation, or the current one has lasted for a while, Bard will have forgotten certain details that you asked him to take into account.
Let’s say you are looking to discover a new recipe and you don’t eat meat, you will have to systematically remind him of this information each time you ask him the question. Currently, Bard can only use your location data or your Gmail inbox, if you agree to give it access, to personalize its responses. But that will soon change with Memory, Bard’s next upcoming feature.
Bard will soon remember you as a true friend
Memory allows users to enter various information about themselves for Bard to remember. In the examples spotted by our colleagues at 9To5Google, we find “I’m trying to eat less meat”, “I have two children” or “formulate shorter answers”. These sentences are, in a way, parameters that will allow Bard to sharpen its responses by personalizing them to the specific needs of each user.
On the same subject — Bard: shared conversations are visible on Google, be careful what you write
By clicking on the Memory tab, it is possible to manage all of these Bard memories by adding new ones or removing those already configured. Note that you can also deactivate the functionality at will, which will not only allow you to start with a blank page if necessary, but also protect your confidentiality if necessary.
Source: 9to5Google