Apple’s New AI Photo Tools Blur the Line Between Record and Creation

The iOS 27 developer beta brings three primary AI features to the iPhone, fundamentally changing its native editing capabilities. The Clean Up tool, previously limited to on-device processing, now leverages cloud models to remove unwanted subjects with far greater precision, effectively matching the performance of Google’s Magic Editor. Similarly, the new Extend feature allows users to widen frames by generating plausible edge content. It generally favors symmetry and avoids aggressive manipulation, though it occasionally introduces minor, artificial elements like unprompted houseplants into the background.

Spatial Reframing proves more contentious. Designed to simulate physical camera movement to correct composition after the fact, the tool attempts to shift perspective in a three-dimensional space. While useful for minor adjustments, the feature struggles with complexity. When tested on a photo of Apple executives, the AI hallucinated a person who was not originally present. Furthermore, using the tool on close-up portraits often results in an uncanny valley effect, skewing facial features as the software struggles to fill in missing data. Although Apple attaches Synth ID labels to these modified images, these disclosures are often buried within sub-menus, doing little to address the broader erosion of photographic trust. By enabling users to alter the provenance of a scene, Apple is subtly normalizing the idea that a photo is no longer a fixed record of a moment, but a malleable draft.