Personal intelligence, a feature that connects user data across Google products to provide personalized answers, is deployed in Vietnam.
In an announcement on April 15, Google said that Vietnam is one of the next markets to deploy Personal Intelligence, after the feature was available in the US earlier this year.
The feature is in the Gemini chatbot, connecting information from many Google applications that users are using such as Gmail, Google Photos, Search. According to Google, this helps turn Gemini into a “dedicated assistant” when connecting data from many individual systems.
For example, when users want to go to Da Lat, they may have ticket booking information in Gmail, have photos of the intended destination in Photos, watch experience videos on YouTube, and search for information on Search. Users can now ask Gemini to “summarize my plan to go to Da Lat”.
Google Gemini AI application logo on the phone. Image: Luu Quy
According to Google, this chatbot will then not only perform a regular search command, but will be able to synthesize reservation information into a full itinerary, suggest restaurants based on videos about local culinary trends that the user has watched, and even find screenshots of local maps or photos of a gift that the user saved weeks ago.
To activate, users can open Gemini, go to Settings and help, select Connectable applications.
The search giant also confirmed that users have complete control over which apps to link and that the setup process is designed “to ensure absolute simplicity and security”.
According to Google, its personal intelligence possesses two core strengths: the ability to reason on complex data sources; and retrieve specific details. For example, from an email or photo, the system can answer a user’s question. Gemini often combines both of these capabilities, processing text, images and video seamlessly to provide personalized answers to each user. In the response, the chatbot will also cite or explain which connected sources the information it used is from for the user to verify when needed. However, Google also admitted that this feature “sometimes makes mistakes” because it may not properly understand special or diverse contexts from users.
In addition, the company also confirmed that it has set up barriers with sensitive data, and data from Gmail mailboxes or Google Photos libraries will not be used to train the model.
“Our goal is to improve your experience while keeping your data secure and completely in your control,” the announcement reads.
The interface selects applications that can connect to Gemini. Screenshot
Personal intelligence on Gemini will be rolled out in Vietnam starting with accounts subscribed to eligible AI Plus, Pro and Ultra plans and “aims to bring this feature to free users in the coming weeks”. Once enabled, this feature will work across the Web, Android and iOS platforms and with all models in the Gemini model chooser.