Google just made the most ambitious land grab for personal data in AI history — and it’s asking you to opt in.
Gemini’s Personal Intelligence feature, which allows the AI assistant to pull information from your Gmail, Google Photos, YouTube watch history, and Search data to deliver personalized responses, is now rolling out globally. The expansion, reported by The Verge on April 14, brings the feature to users outside the United States for the first time — everywhere except the UK, Switzerland, and the European Economic Area.
The feature launches first for Google AI Plus, Pro, and Ultra subscribers before eventually reaching free users. It’s the clearest signal yet that the battle for AI dominance is shifting from whose model is smartest to whose model knows you best.
What Personal Intelligence Actually Does
The feature, which launched in beta in January 2026, goes beyond the contextual awareness that Gemini previously offered. Where older versions could reference your past conversations, Personal Intelligence can retrieve specific details from text, photos, or videos across your Google apps.
Google gives this example: you need new tires for your minivan, and you’re standing at the tire shop. Any chatbot can look up tire specs. But Gemini with Personal Intelligence goes further — it suggests options based on your family road trips (found in Google Photos), pulls your license plate from a picture in Photos, identifies your car’s trim by searching Gmail, and cross-references ratings and prices.
The feature is opt-in, and Google emphasizes that you can choose which specific apps to connect. You can correct Gemini about your preferences at any time and regenerate responses without personalization. There’s also a “For you” chip on the homepage that offers personalized prompt suggestions.
The Privacy Architecture — and Its Gaps
Google’s privacy pitch is built on two pillars: the feature is off by default, and your data doesn’t leave Google’s ecosystem.
“Your personal data already lives at Google securely,” the company argues, framing Personal Intelligence as a natural extension of services you already use. “We don’t need to send it elsewhere to start personalizing your experience.”
On training, Google draws a distinction between using your data to respond to you and using it to train models. “We don’t train our systems to learn your license plate number,” the company explains. “We train them to understand that when you ask for one, we can locate it.”
The training process uses “limited info, like specific prompts in Gemini and the model’s responses,” after filtering or obfuscating personal data from conversations.
The gaps in this architecture are worth noting. Google acknowledges “inaccurate responses” and “over-personalization” — where the model makes connections between unrelated topics — as known issues. The company specifically calls out struggles with “timing and nuance, particularly regarding relationship changes, like divorces, or your various interests.”
As Google’s own example illustrates: seeing hundreds of photos of you at a golf course might lead Gemini to assume you love golf. But you don’t love golf. You love your son, and that’s why you’re there.
The Competitive Context
Personal Intelligence isn’t happening in a vacuum. It’s Google’s answer to a question that every major AI company is now asking: how do you make AI genuinely useful rather than merely capable?
OpenAI has been expanding ChatGPT’s memory and personalization features. Apple is building contextual awareness into Siri with Apple Intelligence. Microsoft is weaving Copilot deeper into Office and Windows. But none of those competitors have what Google has: the largest collection of personal data on the planet.
Gmail has 1.8 billion users. Google Photos holds trillions of images. YouTube processes over 500 hours of video uploaded every minute. Google Search handles 8.5 billion queries per day. When Gemini connects all of those, it’s not just personalizing responses. It’s building a model of your life that no competitor can replicate.
The exclusion of the UK, Switzerland, and the EEA isn’t random. It’s a direct consequence of the EU’s General Data Protection Regulation and the UK’s post-Brexit data protection regime. The same privacy laws that European policymakers designed to protect citizens from exactly this kind of data aggregation are the laws preventing Google from offering the feature there.
What Changes When AI Knows You
The shift from general AI to personal AI is more consequential than it sounds.
A general AI is a tool. You ask it questions, it gives you answers, and the interaction ends. A personal AI is something different. It learns your habits, preferences, relationships, finances, health concerns, and daily patterns. It becomes an extension of your memory and decision-making.
That’s powerful. It’s also a dependency. Once you start relying on an AI that understands your life context, switching to a competitor means starting from zero. Google isn’t just making Gemini more useful. It’s making Gemini harder to leave.
The commercial logic is impeccable. The privacy implications are serious. The societal question — whether any single company should hold this much intimate data about this many people — is one that regulators, users, and Google itself are still working through.
For now, the feature is opt-in. But Google’s track record with opt-in features that eventually become opt-out, or that become so useful that opting out feels like a downgrade, is well documented.
Personal Intelligence is available now on Android, iOS, and web for AI Plus, Pro, and Ultra subscribers outside the UK, Switzerland, and the EEA. Free tier access is coming soon.
Sources
- The Verge — “Gemini with Personal Intelligence is rolling out in more regions”
- 9to5Google — “Gemini rolling out ‘Personal Intelligence’ beta that uses Gmail & your Google apps data”
- Google Blog — “Personal Intelligence: Connecting Gemini to Google apps”
