Did you know Google Prepares For New Gemini Personalization Model That Accesses Users’ Search History
Google is working on a new personalized Gemini model that can go through
users’ search history to provide them more detailed and customized
responses.
As shared by tech media outlet Android Authority,
a new investigation proved how the model is in the pipeline and getting
prepared to attain access to users’s searching history. This will give
users exactly what they’re looking for based on the search results.
While
the Personalization option on Gemini isn’t up for grabs to users right
now, Google is certainly making it available as the newest beta version
for the Android app. This means it would appear alongside the usual
Gemini chatbot.
The new personalization feature is quite different from Gemini’s other
variants. It requests permission to read the entire Google Search
history right before the user begins. You will similarly need to switch
on Web and App Activity settings for it to work.
Once it provides
Gemini access to search history, you get the chance to ask queries
depending on your previous use of Google Search. As shown by the report,
it displays Gemini giving more detailed replies to queries like display
my search history and when did I first search about the Gemini chatbot.
The privacy issues here are going to be obvious. Search histories contain a lot of details about a user. Let’s not forget how modern day AI tools increase the ease with which personal data gets exploited.
The Android maker has taken serious steps to address some concerns on this front. They are messaging users about how they can disconnect the search history from Gemini if they wish to do so. Google also shared that chats with such models aren’t used to better Gemini. Similarly, they shared that nothing is stored outside the chat history. As a whole, the chats get deleted from the Gemini apps after nearly two months.
However, we must remember that outsiders can attain access to the Gemini app and therefore end up misusing it. So the risk is high. Nevertheless, we can see such tools rising in usage in places such as law enforcement and intelligence agencies. For now, this is just a taste of what might be coming soon in the future.