On April 14, Google rolled out Gemini's Personal Intelligence feature to paid users worldwide. The feature connects the AI assistant with nearly all personal data in a user's Google account — including Gmail, Calendar, Photos, YouTube, and Maps.
Europe is not included.
What Data It Can Read
Personal Intelligence accesses a broad range of data:
- Gmail, Calendar, Drive, and other Workspace apps
- Google Photos (all photos and videos)
- YouTube (favorites, watch history)
- Google Maps (location history, saved places)
- Google Search history
- Other first-party Google apps
With this data integration, Gemini no longer needs users to explain context each time. It knows which hotel you booked last week, what video you are watching, and your shopping habits.
Google provided several use cases:
- Recommending clothes based on past purchases and style preferences
- Using shopping receipts to help troubleshoot device issues
- Planning trips based on dietary restrictions and booked flights
- Discovering new interests based on reading history
Privacy: Google's Explanation, Users' Concerns
Anticipating controversy, Google issued a statement ahead of the launch: Gemini will not use your Gmail data to train AI models. It uses the data to answer questions, not to feed training sets.
Technically, this distinction is real — runtime retrieval-augmented generation and pre-training datasets are different. But from a user perspective, it relies on trust in Google's promises, with no third-party independent verification.
Control mechanisms:
- Opt-in feature, enabled by default after initial setup
- Toggle in the tools menu, can be temporarily disabled per conversation
- Users can choose which apps to open or block
Who Can Use It, Who Is Excluded
Available immediately from April 14:
- Global Google AI Plus, Pro, and Ultra paid subscribers
Rolling out over the coming weeks:
- Free-tier Gemini users (globally)
Explicitly excluded:
- European Economic Area (EEA)
- Switzerland
- United Kingdom
Europe's absence is due to GDPR. Integrating multiple first-party data sources into an AI system requires full compliance justification under European regulatory frameworks, a hurdle that has not yet been cleared.
Subscription pricing: Google AI Pro at $19.99/month, Ultra at $249.99/month. Ultra users also get the Project Mariner browser agent (handling up to 10 tasks simultaneously) and the Project Genie interactive world generation feature.
Divergence from Apple's Approach
Apple's Apple Intelligence takes a different path: core computation happens on-device, keeping personal data as local as possible.
Google's choice: integrate cloud data into AI while promising not to train models on user data.
Both approaches have their merits. Google's method is more powerful due to more complete data sources; Apple's method is more privacy-pure because data never leaves the device. Which to choose is up to the user.
But one thing is certain: the shift from AI assistants that answer questions to ones that already know you is irreversible. The global launch of Personal Intelligence is just one milestone on that path.
Sources: CocoLoop, Gemini app starts rolling out Personal Intelligence globally (9to5Google); What Gemini features you get with Google AI Plus, Pro & Ultra (9to5Google); Google is expanding Personal Intelligence to Gemini users globally and it's a huge shift (Digital Trends)