Google is turning Search into something that feels less like a search box and more like a personal assistant that actually knows you — because now, it kind of does. With the new Personal Intelligence update, AI Mode in Google Search can tap into your Gmail and Google Photos to tailor results and answers around your real life, not just your queries.

On paper, the idea is straightforward: instead of you telling Google your preferences over and over, AI Mode quietly picks them up from the data you already have sitting in Google’s ecosystem. Hotel bookings in your inbox, old trip photos, receipts, screenshots, random confirmations — all of that becomes context for more “you-shaped” answers. You’re not just searching “things to do in Goa,” you’re asking something that can be answered with an understanding of what you like, where you’ve been, and what your calendar and photos say about your life.

Take travel planning, which is one of Google’s favorite examples. If you ask AI Mode to plan a weekend in New York, it can look at your Gmail to find your flight details and hotel booking, then peek at your Photos to see the kind of trips you’ve taken before. Maybe your gallery is full of street food stalls, flea markets, and sunsets rather than fine dining and museums. The system can then pitch an itinerary built around hole-in-the-wall food joints, rooftop spots, and waterfront walks, instead of a generic “Top 10 things to do in NYC” list you’ve seen a hundred times.

The same logic applies to smaller, everyday asks. If you’ve got years of brunch and dessert selfies on Google Photos, AI Mode can lean into that when you ask for “places to eat nearby,” surfacing ice cream parlors or dessert cafés that match the vibe it’s inferred from your history. When you’re shopping, it can prioritize brands you’ve actually bought before, using order confirmations and receipts buried in your Gmail as a quiet signal of what you already trust. In theory, that’s less scrolling past irrelevant options and more “oh yeah, I’d probably buy that.”

What’s new here isn’t just personalization — Google’s been personalizing Search for years — but how deep the personalization goes and how explicit the AI layer is about it. AI Mode behaves more like a chatty assistant sitting on top of your Google life, pulling in context live from multiple apps instead of relying purely on your past searches or ad profile. This builds on the same Personal Intelligence feature that rolled out to the Gemini app earlier, where Gemini can already connect to Gmail, Photos, YouTube history, and Search to answer deeply specific questions like “what’s the gate code my landlord emailed me last year?” or “what was the book my friend recommended in that email?”

At least for now, this extra-brain mode isn’t for everyone by default. Google is limiting the AI Mode + Personal Intelligence combo to paying Google AI Pro and AI Ultra subscribers in the US, and only on personal Google accounts. It lives under the Labs umbrella, which is Google’s way of saying “this is experimental, don’t be surprised if it’s rough around the edges.” You have to explicitly opt in and connect Gmail and Photos; nothing gets turned on silently in the background, and you can disconnect those apps whenever you want.

The privacy story is where most people are going to hesitate, and Google clearly knows that. The company keeps repeating the same line: AI Mode is “built with privacy in mind” and “doesn’t train directly” on your Gmail inbox or Photos library. Instead, the training is limited to specific prompts you type into AI Mode and the model’s responses, which are used to improve how the system behaves over time. In other words, your broader email archive and photo library aren’t themselves dumped into some central training set, but your interactions with the AI are still fair game for tuning the model.

From a user perspective, that distinction will either feel reassuring or like technical hair-splitting. You’re still giving Google permission to algorithmically parse personal emails and images in real time to generate answers. For a lot of people, that’s going to trigger a “this feels too intimate” reflex, even if the connections are local to your account and wrapped in access controls. And it doesn’t help that we’re all coming off a decade of tech companies promising privacy “by design” while quietly expanding how they use data.

Google is trying to counter that skepticism with a few guardrails. Personal Intelligence is off by default; you choose which apps to connect, and you can revoke access at any time from settings. The company says there are limits on how far back the system can reach and notes that deleted emails and photos aren’t accessible to Personal Intelligence. There’s also an acknowledgment baked into the messaging that this thing will mess up: it may misinterpret your interests, latch onto outdated information (like an ex’s name or a hobby you dropped years ago), or draw weird conclusions from edge-case data.

When that happens, you’re expected to nudge it back in line. If AI Mode surfaces a bizarre recommendation or just gets you wrong, you can correct it in the chat or give it a thumbs down, which feeds back into the system. That essentially turns you into a human editor for your own AI assistant, teaching it what parts of your digital footprint are actually meaningful versus accidental noise. It’s a glimpse of what everyday AI use is probably going to look like: less “magic box that just knows” and more “helpful, but needs regular supervision.”

Zoom out, and this move is also a competitive flex. Google has something rivals covet: billions of users who already live in Gmail, Photos, Search, Maps, and YouTube. Personal Intelligence is a way to weaponize that integration in the AI race, promising answers that OpenAI, Microsoft, or anyone else can’t easily match because they don’t have that same first-party data at scale. It’s an attempt to make staying in the Google ecosystem feel not just convenient, but functionally better, almost like you’re leaving capabilities on the table if you move elsewhere.

Of course, that same strength is exactly what makes privacy advocates nervous. When a single company can read your email, recognize your face in photos, follow your watch history, and now fuse all of that into a single AI layer, the question isn’t just “is this useful?” but “how concentrated is this power?” We’ve already seen earlier flare-ups like the “Scroogled” era debates around Gmail scanning ads, and while the mechanics have evolved, the core tension is the same: personalization versus surveillance, convenience versus creepiness.

For everyday users, Google’s pitch is very simple: imagine Search that already knows the confirmation code you can’t find, the trip you forgot to finish planning, the brands you genuinely buy, the kind of weekends you actually enjoy. If that sounds like a lifesaver, Personal Intelligence inside AI Mode is built for you — as long as you’re comfortable letting an algorithm sit that close to your inbox and camera roll. If it doesn’t, you can leave it off and keep using Search the old way, but you’ll be watching from the sidelines while Google tries to redefine what “just Google it” means in an AI-first world.