Gemini Intelligence is the future of Android, and I already hate it

Gemini has been a big part of Android for the last couple of years. Google hasn’t been quiet or shy about its ambitions for the AI, regularly rolling out new models and features and finding new ways to integrate Gemini deeper and deeper into the Android OS.

Google is pitching Gemini Intelligence as a new suite of more advanced AI capabilities for Android phones — including task automation with your apps, Chrome auto-browse, and even UI tweaks to better emphasize Gemini’s presence. It’s Google’s biggest push yet to make Gemini the centerpiece of Android, and I think it’s a gigantic mistake

The main thing Google is hammering home with Gemini Intelligence is that it wants Gemini to do more things for you. You probably use Gemini now to ask about the weather or add things to your calendar, but Google is envisioning a future where you rely on Gemini to go a step further and do more things on your behalf across your apps and web browser.

Google shared a few examples of this, such as having Gemini book a last-minute fitness class for you or buy concert tickets. In one demo, the company showed someone taking a picture of a travel brochure, asking Gemini to book a tour similar to it on Expedia, and Gemini seamlessly finding and booking one in the background. Google also shared a few examples of how Gemini will be able to do things for you through its Chrome auto-browse feature. Google says you’ll be able to have Gemini browse the web on your behalf to find parking for an event, track out-of-stock items, and more. The goal is for Gemini to use your apps and web browser to handle tasks on your behalf so you don’t have to.

That’s ultimately true of the other capabilities Google is touting for Gemini Intelligence. The company claims that an improved autofill feature will intelligently add information such as your passport number and license plate to related forms. The new “Rambler” feature in Gboard is supposed to provide a much more powerful voice-to-text experience — one that can change your text’s formatting and style in ways that just aren’t possible today.

Taking Google at its word, all of this sounds impressive. Having your phone’s AI be more proactive and perform tasks on your behalf seems great. However, it’s one thing to say that Gemini will be able to do all these wonderful things, and another entirely to actually implement such ambitious features.

And to be perfectly honest, I don’t trust Google to nail the execution.

As it stands today, Gemini — like every AI — isn’t always reliable. When asking a basic question, Gemini still hallucinates incorrect answers. When I’ve used Gemini to generate an image or edit a photo, even with extremely detailed instructions, it regularly misinterprets my guidelines and gets completely off-base. Now, Google is asking me to trust that same AI to take greater control of my phone with even more complex requests.

Beyond my concerns with accuracy, I also have my doubts from a technical perspective. Last year, Google hailed Magic Cue as the next great AI feature for the Pixel 10 series. Magic Cue was supposed to interpret text messages, phone calls, and more to surface relevant information when you needed it. As someone who used a Pixel 10 Pro for months after its release, Magic Cue rarely reared its head. And when it did, the suggestions were so barebones and unhelpful that I’d rather the feature just didn’t exist.

That’s not to say AI models aren’t improving, but what Google is touting with Gemini Intelligence is far more advanced and complicated than just about anything else we’ve seen from the AI so far. Considering Gemini’s general repeated failings and how Magic Cue was executed, Google has given me no reason to believe that Gemini Intelligence will work as magically as it claims.

But let’s say Gemini Intelligence works flawlessly. Even if it does everything exactly as Google says it will, I still don’t think it’s an exciting or interesting path forward for Android as a whole.

Google’s Material 3 Expressive was such a bright spot in the tech world last year. While so much of the industry was focused on AI, AI, and more AI, seeing Google redesign Android’s entire user interface with artistic, expressive, and whimsical UI elements felt like the perfect antidote to what so many other companies were focused on (even other parts of Google itself).

Although Material 3 Expressive isn’t going away, seeing Google shift its focus from that to exclusively talking about Android in the context of AI and Gemini this year is supremely disappointing. Google also launched plenty of AI features last year, but that wasn’t the core focus. Android 16 was all about making Android more thoughtful and personal while also including a smattering of AI features. And from what we’ve seen of Android 17 so far, it looked like Google was continuing that mission.

But the way Google is talking about Gemini Intelligence, it feels like the company’s priorities have completely shifted. Material 3 Expressive is still there, but it’s now garnished with Gemini-specific UI elements to make the AI more prevalent. All of Google’s Android-related announcements from this year’s Android Show were squarely in the context of Gemini Intelligence. Hell, Google is now referring to Android as an “intelligence platform.”

Google is being very clear about its intentions for Android, and at the very least, I can appreciate the company being so upfront about them. But as a longtime Android fan and someone who really liked the company’s direction last year, this next chapter of Android seems like a nightmare.