Google is developing a feature that allows users to add preferences to Bard that the chatbot must follow. A sort of “memory” that will allow the AI to respond in an ever more relevant way.
Avoid answers that are not entirely suited to your situation and make exchanges with Bard more fluid. This is the objective of Google, which is currently working to add a “memory” functionality to its chatbot, allowing users to enter preferences to which the AI must subsequently comply in its responses.
The idea? Avoid starting from a completely blank page when you start a new conversation with Google Bard, and thus arrive at more adapted and personalized responses much more quickly – unlike the current situation. Because despite some progress (Bard can refine its responses based on your geographic location, or the content of your Gmail inbox, if you allow it), the personalization of responses remains quite succinct for the moment.
Bard soon to be more insightful?
To progress in this area, Google will soon focus on adding memories to Bard. A rather poetic way of saying that you will be able to add prerequisites to Google Bard’s responses, as demonstrated by the screenshots unearthed by 9to5Google. We discover that this new feature will allow you to “Create a memory” so that Bard remembers certain important information from the start.
In the case of a request for cooking recipes, you can, for example, force Bard to remember that you avoid eating meat, that you are allergic to this or that ingredient, or that you cook for yourself and your children. Perfect for directly getting more relevant meal ideas, with the right portions as a bonus. It will also be possible to force Bard to give short, probably more concise, answers.
Note that this type of personalized instructions will not be exclusive to Google Bard. As 9to5Google points out, ChatGPT has already implemented a similar system.
Finally, let us point out that for reasons of confidentiality, or even practicality, it will be possible to quickly deactivate Bard’s memories using a button located on the left side of the screen. Useful for demonstrating to a friend without Bard taking your preferences into account… or if you want to question Bard on topics that you would prefer him to forget for the next session. Which also suggests that the acquisition of memories on Google’s AI will be partly proactive.