Apple is adding new generative AI tools to the Photos app with iOS 27. Jon McCormack, head of the iPhone camera team, stated that this move is not about showcasing AI for its own sake like competitors, but about making photo editing more controlled and user-friendly.
The WWDC presentation highlighted new features including an improved version of the Clean Up tool, along with two new tools called Extend and Spatial Reframe. These tools expand the photo’s surroundings or change the perspective of the image, filling in missing areas artificially as needed.
How the New Tools Work
Apple’s approach is to design these features to correct composition errors noticed after taking a shot. McCormack explained that the goal is to enable users to make simple edits without needing Photoshop-level skills, providing powerful flexibility to everyday users.
The company does not allow unlimited image generation with these tools. The Clean Up tool cannot remove the main subject of a photo, the Extend feature can only be used once per image, and it can expand the image by up to 25%.
It was noted that any artificially generated pixels will remain only in the background, and altering the main subject’s face is not planned. Product manager Della Huff said the models are trained to avoid unnecessary fabrication and to exclude details that are not essential for the purpose.
However, in a developer preview test, the Extend feature was observed to generate new tables and people who do not exist in real life when expanding a scene with people in the background. Huff explained that the feature tries to match the existing aesthetic, otherwise the extension could look artificial.
Usage Limits in Daily Life
Apple will impose daily usage limits on these new camera tools but has not disclosed the exact numbers. It was mentioned that users will need an iCloud subscription to use Extend, Spatial Reframe, and Clean Up features multiple times per day.
Another key theme of this year’s presentation was natural language processing. Examples included freer conversations with Siri, creating events in Calendar, defining commands in the Shortcuts app, and describing extensions in Safari; however, photo editing was emphasized as not being delegated to Siri.
Huff said Siri could assist with minor touch-ups but that more open-ended tasks like perspective changes would not provide a good user experience. McCormack added that the main goal of integrating Siri into the camera app is to reduce friction, and that the Visual Intelligence function’s camera-centric design fits this area more naturally.
Apple’s approach is deliberately distinct from another AI product, Image Playground. While the Photos app is positioned as a safe space to preserve memories, Image Playground is designed for freer, more creative use.
McCormack argued that a photo should represent a genuinely lived moment, and Apple’s goal is to develop tools that maintain the integrity of that moment. The company also plans to add Google DeepMind’s SynthID technology later this year to provide invisible watermarking for edited images.