Google has announced a significant upgrade to its Gemini artificial intelligence platform, introducing a new capability called Personal Intelligence that enables deeper, more personalized assistance by securely accessing a user’s broader Google ecosystem.
The new feature allows Gemini to utilize information from Gmail, Google Photos, YouTube, and Search to respond more effectively to user queries. While Gemini has previously been able to reference app-based information, the latest update marks a major advancement by enabling the AI to reason across multiple, complex data sources simultaneously.
According to Google, Gemini can now extract specific details from emails, images, and other personal data — and even combine those insights — to generate “uniquely tailored” responses.
Explaining the feature in a blog post, Josh Woodward, Vice President of Google Labs and Google Gemini, shared a real-world example. While standing in line at a store, he used Gemini to help purchase tires for his minivan. The AI identified the vehicle’s license plate number from an image in Google Photos, retrieved vehicle specifications from Gmail, and recommended suitable all-weather tires based on additional photos from a family road trip.
Google emphasized that Personal Intelligence is designed with user choice and privacy safeguards. The feature includes sourcing transparency and also allows users to request responses without any personalization. Guardrails have been implemented around sensitive areas such as healthcare, the company said.
Acknowledging privacy concerns, Woodward clarified that connecting personal apps is disabled by default.
“Users choose whether to turn it on, decide which apps to connect, and can turn it off at any time,” he said.
He further explained that Gemini is not trained directly on personal user data, but rather on how to respond to specific prompts.
“We don’t train our systems to learn your license plate number; we train them to understand that when you ask for one, we can locate it,” he noted.
As with other agentic AI systems, Google acknowledged the potential for errors, including unintended connections between unrelated data or challenges in interpreting nuance and context. To help improve accuracy, users are encouraged to provide feedback using a thumbs-down option.
The Personal Intelligence feature is currently available in beta for eligible Google AI Pro and AI Ultra subscribers in the United States. Google said expansion to additional countries and the free tier is planned in the near future.
The tool works across the web, Android, and iOS platforms, supports all models available in the Gemini model picker, and is limited to personal Google accounts. Google also confirmed that enhanced functionality for AI Mode in Search will be rolled out soon.