- By Vikas Yadav
- Sat, 30 Sep 2023 04:09 PM (IST)
- Source:JND
Google Bard, the tech giant's answer to OpenAI's ChatGPT, could reportedly get its own memory. This will allow the chat model to keep track of user preferences while framing responses, according to a 9To5Google. In the current scheme, whenever a user tunes into Bard for conversations, it generates responses without remembering personal preferences, which implies the answers are similar worldwide.
While the model can personalise responses based on the location and extensions, it is tough to remind the model at the start of every question about select conditions to keep in mind while framing the answer. But, this challenge could soon be addressed via the 'Memory' feature that Android app developer Dylan Roussel discovered.
The new feature will let Bard remember certain conditions set by a user and learn specifics to enhance future responses. In the screenshot shared in the report, the new option is visible. Bard can be told to keep in mind things like -"I try to avoid eating meat," I have 2 kids," and "Please give shorter responses."
The 'Memory' feature in Bard was discovered by Android app developer Dylan Roussel. (Image:Unsplash)
The memory options shall help refine responses in the future, like offering recipe suggestions, making travel plans and more. ChatGPT has a similar feature to memory called "custom instructions." In Google's chat model, this page will likely be accessible from the left side of the menu, where users will be able to enter new preferences. An option to delete custom memories will likely also be available.
An option to toggle the function off will also be available in the same window, according to the report. This would enable the model to function well during demonstrations and preserve privacy. On a related note, as the AI race between the big tech companies continues to heat up, OpenAI recently introduced fresh enhancements for its model and claimed ChatGPT can "now see, hear, and speak."