Apple's camera chief on iOS 27 AI: 'superpowers,' not hype

With iOS 27, Apple has done something people both wanted and feared: generative AI is now built directly into the camera and the Photos app. The phone can paint in pixels that never existed in the original shot.

What the new Photos app can do

The updated Photos app gets a set of generative tools that work on top of an already captured image. The principle is the same everywhere: the algorithm analyzes the context of the frame and creates new content that blends visually into the original.

The Generative Expand function pushes out the borders of a shot: the AI paints in more sky, continues the background and adds space where the camera did not capture it. Clean Up removes unwanted objects such as a random passerby, a pole or a trash bin, and the algorithm rebuilds the background as if the obstruction had never been there.

John McCormack, who oversees the development of camera systems at Apple, explains the company's approach through a simple question: does the feature solve a real problem for the user? If not, it will not make it into the iPhone.

"We don't add AI for the sake of AI," John McCormack, Wired.

"Superpowers" is not a metaphor

"Superpowers" is not a marketing slogan but a working term inside McCormack's team. They use it for tools that make possible what once required Photoshop and hours of work, or was not achievable at all.

Apple is betting on personal use cases: family photos, travel, portraits. In these cases, removing a stranger from a frame or extending a photo to fit a wallpaper format is a clear and useful operation, not manipulation.

Generative features of the new Photos in iOS 27:

• Generative Expand: extends the frame and paints in the background
• Clean Up: removes objects and fills the space intelligently
• Restoration of blurred and noisy details in a shot
• Automatic suggestions for improving composition
• Adapting a frame to Stories, wallpaper formats and 4:3 ratios

According to McCormack, every feature went through an internal check: does it bring real benefit to a specific person, or does it merely show off what the technology can do?

Pixels that were not there

This is where the uncomfortable conversation begins. Technically, a photo with added or altered pixels is no longer an accurate document of reality. The photography industry has long argued over where retouching ends and falsification begins.

Sony, Nikon, Canon and large media platforms are already adopting the C2PA standard, a digital signature technology that records the full history of an image and explicitly labels generative changes. This matters for news and documentary photography, where authenticity is critical.

Apple has not yet publicly stated that iOS 27 supports C2PA. McCormack points out that the generative features were designed for everyday scenarios, not journalistic ones. But the line between a "family photo" and an "archival document" is blurry, especially now that the same images end up in courts, insurance companies and visa applications. Meanwhile, the iPhone does not explicitly warn that some of the pixels in a frame were generated by an algorithm rather than captured by the lens.

What this means

Apple is officially entering the era of the generative camera, where Google Pixel has been heading for the past few years with Portrait Unblur and Magic Eraser. The difference is that the iPhone is the most widely used camera in the world, and Apple's decisions automatically become the new standard for hundreds of millions of users.

The "not for AI's sake" stance is the right signal, but the real test begins in the fall. That is when we will see whether Apple is ready to explain to people what in their frame is reality and what is the work of a neural network.