Apple’s Director of Camera Technology Believes AI Can Enhance Your Abilities

What even is a photograph in today’s world?

As major tech companies integrate generative AI into their camera software, the distinction between authentic images and manipulated ones continues to fade. For instance, Google and Samsung phones now offer features that allow users to significantly modify photos by removing individuals, shifting people within frames, and even inserting new elements into images.

Apple is joining the trend with new generative capabilities in its Photos app. However, Jon McCormack, head of the iPhone camera division, emphasizes that Apple is taking a more cautious route compared to competitors, insisting that the company isn’t “doing AI for the sake of AI.”

During its annual Worldwide Developer Conference on Monday, Apple showcased various AI features that will be integrated into the Photos app in iOS 27, set to launch on iPhones later this year.

While the iPhone Photos app already includes a Clean Up tool for removing unwanted objects, enhancements in iOS 27, driven by Apple’s refined AI models, will improve its functionality. The update introduces two new features—Extend and Spatial Reframe—that allow users to enlarge the area around a photo or adjust the perspective, generating artificial pixels in the process. The camera anticipates what should fill these areas and creates it accordingly.

McCormack highlights that there’s a significant backlog of challenges that AI is now addressing, and these fresh features are intentionally designed. “You don’t have to understand every intricate detail of Photoshop or another program—it empowers everyday users with remarkable abilities,” he explains.

However, Apple aims to prevent users from excessively manipulating their images and creating unrealistic alterations. (At least within the Photos app; various tools in the App Store enable more extensive photo creation.) The generated pixels will only modify the background, leaving the primary subject’s face unchanged. For instance, with Clean Up, you cannot delete the main focal point in the image. The Extend feature can be used only once, expanding the image by 25 percent; you can’t save and edit it repeatedly to stretch it further with AI.

McCormack also mentions that Apple plans to incorporate Google DeepMind’s SynthID technology later this year to implement an invisible watermark indicating that these images have been modified using generative AI. Platforms where users share these photos may flag them as AI-edited. (Just keep in mind that research indicates digital watermarks aren’t entirely reliable.)

“A photograph represents something that truly occurred,” McCormack states. “We believe strongly in the idea of authentic storytelling in your life—when you capture images, you’re preserving memories, encoding moments of your life that you can revisit. It’s essential for us to develop tools that maintain the integrity of those moments.”