Schrödinger sock. That’s how I describe the child’s foot apparel that was not visible but brought to photo-realistic life by Apple Intelligence running on iOS 27 Dev Beta. The sock exists in the liminal space between assumption and reality. Only generative AI can make it real, even if the child in the photo was never wearing his other sock. Such is the power of the upcoming update to Apple’s iOS and other platforms, and of the up-to-now relatively disappointing Apple Intelligence: Apple takes on Artificial Intelligence so carefully applied that it seemed woefully behind the competition.
But that was before.
Today, after its unveiling on Monday at WWDC 2026, we have the new Apple Intelligence and its powerful image-altering and creation tools that are the product of powerful off-device models built in cooperation with Google, or rather, that incorporate Google’s generative models but apply them with Apple flair.
Image 1 of 4
In the demo I saw, designed to show off the power of Apple Intelligence’s new and enhanced Image Editing tools, we looked at the more powerful clean up that can do something Apple has never tried before in image editing: create something biological out of nothing, or at least infer the existence of something human we clearly cannot see. In this case, a child’s socked foot. In the image (which I sadly cannot share here, but there are other examples above), the boy is seated on the floor, with one leg partially obscured by a large stuffed bear.
As before, we can trace over such distractions like a chair and the bear, and then tap Clean Up to remove them. Removal is something Apple Intelligence is already quite good at, but replacing large image parts or creating what should be there was really not part of Apple’s generative image editing strategy. Now it is.
Like magic, we could see the rest of the child’s leg and his other socked foot. No matter that this small boy might not have been wearing a sock on that foot, it got the same sock. As I said, we’ll never know if that sock ever truly existed, but in the mind’s eye of Apple Intelligence, it does.
Reframing the story
Image 1 of 2
The other tools similarly stretch the bounds of memory and reality. Spatial reframing is a wild Apple Intelligence tool that applies a sort of 3D model to the image, letting you subtly rotate the subject and background this way and that to better frame your image. You can even grab and shift the subject a little bit. I watched them do this with an image of a woman sitting on the grass. I must admit that the final result did look better, and I guess no one will remember the less perfectly composed photos. At least you can always go back to the original, imperfect image to remember how things truly were.
Perhaps the best and least drastic of these generative image alteration features is the new expansion tool that, like AI tools in, say, Adobe Firefly, lets you basically extend the background in a photo to make it a better fit for a certain frame or wallpaper. It’s something I could see myself using quite a bit, especially because it leaves the subjects alone.
Create something new
Image 1 of 3
Yes, friends, we are officially entering uncharted Apple territory. For a company that prides itself on photographic excellence with one of the best cameras, sensors, and image pipelines in the business, it is now fully in bed with photorealistic imagination creation.
In one demo I saw, a young man wanted to create a flyer to promote a, I think, a bake sale for his friend, Phil. As before, Image Playground lets you choose seed photos and subjects from your Photos library, but where the old app would stick to basically comic book renderings of these subjects, the new app, with new Gemini-infused models backing it (they all live on Apple’s Private Cloud compute), happily creates photorealistic images.
It’s now a powerful tool, and I watched as we selected the subject (Phil) and described him piping a cake and with a bowl of fruit nearby. Soon, our buddy Phil appeared on screen looking very much like a master baker. No matter if the guy ever baked a cake in his life, let alone decorated one. He was doing it here. There’s nothing real about this image, and while Apple likes to talk about maintaining the essence of our photos in the Apple Intelligence-powered Photo editing tools, this other land will create full-scale image narratives from whole cloth.
Yes, there is some sonic dissonance here.
Image 1 of 3
Even so, Image Playground is now a vastly more powerful tool. After we gave Phil his fake bakery skills, we decided that the bowl of berries we requested for the image was not the right kind of berries. No matter, we simply used the Apple Pencil to select the berry bowl and then asked the app to switch it to blueberries. We also asked for a cupcake tower because Phil’s baker bonifides were not quite strong enough.
The result was photorealistic perfection.
This is what you wanted, right? We all said Apple was behind. Now it’s pulling even with the rest of the generative image world. It’s progress, though I’m not sure to what end.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
https://cdn.mos.cms.futurecdn.net/GhWhm3VuaaVZNvvmyxDLGa-1920-80.jpg
Source link
lance.ulanoff@futurenet.com (Lance Ulanoff)
