I compared Siri AI to Gemini on Android — and Apple actually understands what most people want from AI Jason Cipriani | usagoldmines.com

I’ve had a love/hate relationship with Siri since the iPhone 4S in 2011. At first, I loved it. I’d frequently ask it for random facts, to tell me a joke, to check the weather, or to send a message. I used it so much after its debut that my wife and kids started calling Siri my girlfriend.

Then the novelty wore off, and my interactions with Siri all but stopped. I didn’t hate Siri; I just no longer found it a tool I wanted or needed to use regularly.

Since then, every time Apple announced new features or capabilities, I’d attempt to reconnect with my old fling, only to grow bored once again.

In June, Apple unveiled Siri AI — a new and improved digital assistant the company touted as smarter, more capable, and full of new features. I was, admittedly, very skeptical. The year prior, Apple demonstrated a more powerful version of Siri that it never delivered.

A few days after the announcement, I installed the developer betas on a MacBook Neo, iPad Pro and iPad Mini — three devices I almost never talk to or interact with Siri on. However, I do often press Command+Space for Spotlight for common tasks. Now, instead of Spotlight, the same key combo brings up a text box you can use to chat with Siri. Also, the new Siri includes a dedicated Siri chatbot app.

For the first six weeks of the beta program, that’s how I used Siri — via the old Spotlight shortcut or directly in the Siri app. I’d often ask Siri something, then feed Gemini the same prompt, and compare the results.

Initially, I was puzzled by the lack of an option to pick the AI model I want to use. As someone who uses Gemini daily, I’m used to switching between the latest Flash-lite, Flash, and Pro models, plus the Extended version for, presumably, better answers. Claude has similar options with Fable, Opus, Sonnet, and Haiku, each with a version number after it.

In the Siri app, there aren’t any options at all. You open it and start typing. It’s simple, if not basic.

There’s a list of past conversations, as you’d see for old threads in Messages or Mail, only these include all conversations you had with Siri, regardless of how you interacted with Siri, be it voice or typing.

At first, the simplicity of the Siri app was something I just couldn’t wrap my head around. How do I know if I’m getting the best possible answer for daily questions or more complex tasks like coding a website? I wanted to change models to ensure it was more thorough. But I couldn’t.

I didn’t trust Siri. How could I if I didn’t know how it was approaching our conversation?

It wasn’t until I installed iOS 27 on an iPhone 17 Pro Max that my use of Siri AI went up, and with it Apple’s entire approach finally clicked.

My typical workflow on an iPhone is to swipe down on the screen, type a question or an app name, and continue from there. I put zero thought into a process I do dozens of times a day. Just swipe, type, and a few seconds later I have an answer. Only now, instead of that gesture triggering Spotlight, it’s Siri.

The more I used Siri, the more I began to enjoy not having to make a conscious decision about what model, version, or mode to use. The end result is exactly what I go into every conversation with an AI bot wanting: Ask for something, get an answer.

To be clear, Apple is routing queries behind the scenes across multiple on-device and cloud models — like AFM 3 Core or AFM 3 Cloud Pro — via Private Cloud Compute. But you never see those names. Apple intentionally hides the engine under the hood, making the technical complexity invisible to the person just trying to get an answer.

And for the vast majority of people, that’s perfectly fine.

The real advantage Siri AI has over Google Gemini on Android is that Siri’s baked into iOS. No matter how or where you interact with it, you get the same exact experience, and a record is kept in the Siri app.

For Android users, Gemini’s implementation is not as cohesive.

There’s a dedicated Gemini app, just like Siri. Only the Gemini app has far more tools, features, and settings. But outside that app, you have to intentionally seek out Gemini on an Android phone.

For instance, there’s a Google search bar on every Pixel’s homescreen. Tap on it to use Google Search, where you’ll then get an AI overview as the top result. But that’s not directly using Gemini — you’re using Google Search.

If you want to talk with Gemini, either by holding in the power button or triggering it with Hey Google, you have to go through extra steps to enable Gemini’s voice features (instead of Google Assistant) — it’s not automatic. Even after I enabled the power button to trigger Gemini on a Pixel 10 Pro XL, I kept getting a pop-up saying I couldn’t talk to Gemini until I tapped one more button to enable the mic.

What. A. Process.

While working on this story, I got an email from Google saying Google Assistant has now been replaced by Gemini on all my compatible devices — and while I welcome the change, using Gemini on Android is still a segmented experience.

On the iPhone, none of that is required. Regardless of how or where you access Siri, you get Siri… now with a side of AI.

I think Apple’s minimal approach to the Siri app is actually a feature on its own, and a darn good one. When I first started experimenting with Gemini, it took me weeks to learn the nuances of when to use specific models; then, as soon as I got comfortable, new models were released, and the process started all over again. That cycle continues even now, nearly two years later.

I still believe, however, that there’s a need and place for more control over AI interactions. Some people, especially those who use AI for more advanced tasks like coding, benefit from full control of the model.

But the vast majority of people — many of whom are scared of AI — will notice that Siri got smarter and can do more things, without realizing (or caring) that it’s a full-fledged AI platform behind the results.