I’ve used Apple Intelligence since day one — here’s why I still prefer Gemini

I jumped on the Apple Intelligence train pretty quickly. I mean, can you blame me? I just wanted to see if Apple had any chance of catching up to Google. It seemed — at least at first — as if Apple was readying its own version of just about everything Gemini could do, too. Apple announced plans for an image generator, photo editing tools, the ability to write and rewrite messages, and more, and it was all coming… eventually.

Well, now it’s been here for about a year, and I’ve been using it nearly the entire time. I’ve tried the whole range of Apple Intelligence features multiple times across several iOS updates, and I’m still underwhelmed. I’ll be sticking with Gemini for the foreseeable future, and here’s why.

When Google first introduced its Pixel Studio on the Pixel 9 series, I wasn’t entirely sold. Yes, it opened the door to a good bit of creativity, allowing you to describe what you wanted to see while carefully navigating intellectual property, but it was missing something: people.

So, when Apple announced that its Image Playground (now called Playground) would not only include people, but people from your Photos library, I was intrigued. I had hoped it would look a little less creepy than the example Apple shared during its WWDC 2024 keynote, which felt somewhat lifeless and flat, but it didn’t. My first few samples came out with creepy teeth and dead eyes, and although they looked like my friends, I didn’t really want anyone to see the results.

More recently, though, Apple has given its Playground a refresh, adding new animation styles and some help from ChatGPT to further reimagine your familiar faces. The result of all this is… it’s still creepy. I still don’t like the way my iPhone 17 Pro transforms me and my friends into animated characters, and I find that the assistance from ChatGPT rarely works, instead leaving me with a blank box where my image should be.

Google, on the other hand, has actually improved its Pixel Studio. No, it hasn’t added the ability to use likenesses of people you know, but it has added humans in general. It’s a significant improvement when you consider that I was previously asking for generic terms like ‘wizard’ and ‘knight’, and still having Google refuse my requests. And honestly, I don’t really need to make cartoons out of people I know — it’s way more fun having Google recreate my run club in the style of Wallace & Gromit anyway.

Sticking with the image generation — or at least editing — theme, Apple’s Clean Up remains mediocre at best. Last year, I put it to the test on a few photos following the Amsterdam Marathon and found that Apple’s version of Magic Editor still needed, well, the magic. It accurately identified people that I might want removed, but it encountered major issues while attempting to remove them.

Unfortunately, it hasn’t gotten much better over the last few updates. Yes, Clean Up can remove a ceiling fan against a plain wall or a vase from a table in front of some bricks, but things fall apart as soon as there’s any sort of complication in the background. You encounter wonky artifacts and weirdly warped results almost immediately, and there’s no good way to override Apple’s attempt or have it take a second try.

With Google’s Magic Editor and the Help Me Edit feature, however, the results are much different. At times, Google still struggles slightly with artifacts, but it’s not nearly as noticeable as Apple’s tool. I’ve encountered a few odd shadows left behind after erasing an object, although I can usually work around the issue with a follow-up request.

I also have to give Help Me Edit a bit more credit for going far beyond the basics of erasing and resizing people and objects. It pretty much gives you a full-fledged editing studio to replace the color of the sky, swap out what you’re holding, expand your image, or just about anything else you might dream up. It’s a level of power I probably don’t need on my Pixel 10 Pro, but one that I’m glad to have.

Besides, with as much of a head start as Google had with Magic Eraser, I’d be worried if Apple caught up to its image editing capabilities after just one year.

Apple is also trying its very hardest to catch up to the high bar Google has set in terms of calling features — but it’s not there yet. It saw Google create Hold For Me and Call Screen and decided to wait several years before taking its own crack at them. And, now that they’re here, they’re fine but far, far behind — like Pixel 3 behind.

For some reason, Apple needed to create its Live Voicemail platform before it could work on Call Screening, which meant waiting until iOS 17 to get started. Google, for its part, has offered voicemail transcription since 2009, which is only a gap of about 14 years — not too bad, I guess. At the very least, I’m glad Apple is catching up, but I’ve found that its filters aren’t as strict and that I’ve had an unwanted call or two slip through the net.

To Apple’s credit, its Hold Assist Detection is pretty good. I held on for a good 20 minutes while waiting to schedule an appointment, and was able to bounce around and clean up the kitchen without paying much attention to my phone until it chimed.

This time, I can’t really dismiss Apple’s efforts as second-best — its features are pretty similar to Google’s. That said, I’m sticking with Gemini simply because I think it’s a bit quicker and I find it just a bit more reliable.

Pretty much the last straw in Apple’s underwhelming AI experience is that it hasn’t delivered on any of the things Siri was promised to do. It’s promised everything from on-screen to personal context (which sounds a lot like Google’s Magic Cue), and should eventually allow Siri to work across multiple apps, giving it the power to handle more complicated queries.

All of that sounds great, and it probably is great, but there’s a problem: none of it is real just yet. Apple still describes all three of those features as in development and will be available with future software updates. It’s been a whole year, and there’s been no sign of movement on any of those features. Google? Yes, it announced and launched Magic Cue, added photo and video support to Gemini Live, and has spent its extra time creating better versions of the AI features that have otherwise defined recent OnePlus and Motorola launches.

Oh, and Circle to Search, which I’m counting as a Gemini feature, is still miles ahead of Visual Intelligence. Although launching Apple’s Visual Intelligence is as straightforward as launching Siri, I’ve yet to find a ChatGPT question it can answer, which is a stark contrast to the quick and reliable Gemini Live. If that’s Apple’s vision for an image-based assistant, I have questions about what it’s actually looking at.

Also, I was promised a more conversational Siri, which I assume means it will behave more like Gemini Live. That’s not here yet, either. Well, Siri does sound a little more natural in Apple Maps, which is nice, but as soon as there’s a delay in your directions, she reverts right back to her usual robotic tone.

I cannot for the life of me understand how Apple can make so many lofty claims about its voice assistant and then fail to deliver on all of them, while also leaving its best chance at competing with Google feeling half-baked.