Google has unveiled a new beta feature for the Gemini app designed to enhance how the AI assistant interacts with its users. This development allows Gemini to generate more personalized and proactive responses by integrating seamlessly with various components of your Google ecosystem, including Gmail, Photos, Search, and YouTube.
Previously, while Gemini could access information from these applications, the new feature enables it to synthesize insights across different data sources. For instance, it can now link relevant threads in your emails with videos you’ve previously watched, offering a contextual understanding that eliminates the need for explicit user prompts.
This beta functionality, termed Personal Intelligence, is switched off by default, allowing users the flexibility to selectively connect their Google apps to Gemini. Google acknowledges that privacy is a concern for many users—thus, the AI will only activate Personal Intelligence when it deems it beneficial.
Josh Woodward, VP of the Gemini app at Google Labs, highlighted two primary capabilities of Personal Intelligence: its advanced reasoning skills across multiple data points and its precision in retrieving specific details, whether from emails or photos, to deliver tailored answers. For example, he recounted a scenario at a tire shop where he needed his car’s tire size. Instead of merely providing generic information, Gemini utilized family road trip photos to recommend suitable all-weather tires. In another instance, it successfully retrieved his license plate number from an image stored in Google Photos.
Users can also expect personalized suggestions for books, shows, and travel plans. Woodward noted that during the recent planning for their spring break, Gemini recommended unique experiences—like an overnight train journey—while steering clear of typical tourist destinations by analyzing past trips and family interests.
To maintain user privacy, Google has implemented measures to avoid making unsolicited assumptions regarding sensitive topics, particularly health-related information. While Gemini is equipped to handle such data upon request, it does not directly train on individual user content from Gmail or Google Photos. Instead, the AI learns from specific prompts within the Gemini framework.
Currently, the Personal Intelligence feature is being rolled out to Google AI Pro and AI Ultra subscribers in the U.S., with plans for a broader international release and eventual availability in Gemini’s free tier.
Key points about the new feature include:
- Personal Intelligence: Offers proactive, tailored responses by connecting across multiple Google applications.
- User Control: Switched off by default, giving users the choice on data sharing.
- Enhanced Contextual Understanding: Combines data from various sources to deliver personalized recommendations.
- Privacy Measures: Does not directly train on personal data, ensuring user confidentiality.
To explore the functionality, users can try prompts such as:
- “Help me plan my weekend in [city, e.g., New York] based on my interests.”
- “Recommend documentaries based on my recent queries.”
- “Based on my grocery receipts and watch history, suggest YouTube channels aligned with my cooking style.”
