Gemini introduces Personal Intelligence feature

The best digital assistants don't just know the world, they know you and help you navigate it. On January 14, the Google satisfied one of its users' top requests: you can now personalize the Gemini:, connecting its applications Google with a single click. This feature, which is being launched in version Beta in the USA, is the company's next step in making the Gemini: more personal, forward-thinking and powerful.

Personal Intelligence securely connects information from apps like Gmail and Google Photos , making Gemini uniquely useful. By enabling it, you control exactly which apps you connect, with each dramatically enhancing your experience. No other product connects Gmail , Photos, YouTube , and Search with a single tap, and Google has designed the setup to be simple and completely secure.

See more articles from iGuRu.gr when you search for news on Google.

iGuRu.gr preferred source

How users use it

Personal Intelligence has two key strengths: the ability to process complex sources and the ability to retrieve specific details (for example, from an email or a photo) to answer your question. It often combines these elements, drawing data from text, photos, and videos, to provide uniquely personalized answers.

The integration of applications through Personal Intelligence promises to significantly simplify users’ daily lives. For example, in the case of searching for new tires for a minivan, Gemini can provide a solution even when the user does not know the technical specifications. While any chatbot can identify general tire characteristics, Gemini goes one step further. It suggests personalized options – such as tires for everyday use or for all weather conditions – by referring to previous family trips recorded in Google Photos . At the same time, it automatically collects ratings and prices for each suggestion. The tool’s assistance also extends to practical issues of the moment, such as retrieving a vehicle’s registration number through information found in Gmail and Photos, saving the user from unnecessary waiting or searching.

Gemini also provides great recommendations for books, shows, clothing, and travel. This technology is particularly valuable in the area of ​​travel planning. By analyzing interests and past travel history through Gmail and Photos, Gemini can bypass “tourist traps.” Instead, it suggests alternative experiences, such as a night train ride, and even suggests specific board games for the duration of the trip.

Mode of operation and approach to privacy protection

Personal Intelligence is designed with privacy in mind. App linking is disabled by default: you choose whether to enable it, specify exactly which apps will be connected, and can disable it at any time. 1 ^When enabled, Gemini accesses your data to respond to specific requests. Since this data is already securely stored with Google , there is no need to send sensitive information to third parties, which is a key advantage of this technology.

At the same time, the origin of the answers is completely transparent. Gemini cites sources or explains what information it used from the connected applications, so that the user can verify it. In case this does not happen automatically, the user can ask for more clarification. In addition, if an answer is not accurate, there is the possibility of immediate correction (e.g. “Remember that I prefer a window seat”). There is also the option to recreate answers without personalization for a specific conversation, or to use temporary conversations.

There are also strict restrictions on sensitive matters. Gemini avoids making forward-looking assumptions about sensitive data, such as health issues, although it can discuss relevant information if the user explicitly requests it.

Google 's goal is to improve your experience by keeping your data safe and in your control. Gemini is designed with privacy in mind and is not trained directly from your Gmail inbox or Google Photos library . To improve its capabilities over time, it is trained with only the minimum amount of information possible, such as specific prompts to Gemini and the model's responses. The personal information you connect is intended only for you.

Let’s look at what this means in the minivan example. The photos from the trip, the license plate photo in Google Photos , and the emails in Gmail aren’t used to train the model. They’re referenced only to give you the answer you asked for. Google trains the model using data like your specific commands and responses, after we’ve taken steps to filter or hide personal data from the conversation with Gemini . In short, the company isn’t training its systems to learn your license plate number—it’s training them to understand that, when you ask for it, it can be traced.

You can read more about Google 's approach to privacy here . At any time, you can adjust your settings, disconnect Google apps , or delete your chat history.

User contribution to improving the experience

Although the Beta version of Personal Intelligence has undergone extensive testing to minimize errors, they have not been completely eliminated. Users may encounter inaccurate responses or “over-personalization” phenomena, where the model incorrectly connects unrelated topics. In such cases, providing feedback via the “ thumbs down ” option is crucial to improving the system.

Gemini may also have difficulty with the timing or context of some information, especially when it comes to changes in relationships, such as divorce or the user’s various interests. For example, having hundreds of photos of a golf course might lead the model to conclude that the user loves that sport. However, Gemini may not realize that the user doesn’t love golf, but is accompanying his son because he loves him. If Gemini makes such an incorrect assumption, the user can simply correct it immediately, for example by saying “I don’t like golf.”

These areas, as well as many others, are still the subject of ongoing research and improvement. For a deeper analysis of the methodology, current limitations, and actions being taken to address them, the relevant scientific paper is available.

How to access Staff Intelligence through the Connected Apps

Starting January 14, the new feature will be rolling out to eligible AI Pro and AI Ultra subscribers in the U.S. ² Once enabled, it will be available on the web , on Android and iOS devices , and on all available Gemini models . The feature is rolling out to this limited group of users to collect data and improve the experience, and will soon expand to more countries and the free version of the service. The feature will also be integrated into Search in AI Mode soon . The Beta is currently only supported for personal Google Accounts and not for Workspace for business or education users.

If you don't see a test invitation on the Gemini home screen , you can enable it in Settings by following these instructions:

1. Open Gemini and tap Settings

2. Press Personal Intelligence

3. Select Connected apps ( Gmail , Photos, etc.)

newsbot

Although the press releases will range from very select to rare, I said I'd pass...because sometimes the editors hide.