How Apple's Extend Tool Changes Photo Editing

Apple's iOS 27 developer beta introduced a suite of AI-driven features in the Photos app, but none generated as much early buzz as the Extend tool. Unlike traditional photo expansion that simply stretches or fills with solid colors, Extend uses on-device machine learning to generate plausible new content beyond the original frame. After spending several days testing it on an iPhone 17 across dozens of photos, I can report that this feature marks a significant leap forward for mobile photo editing.

Headline: Extend Tool Delivers Mixed but Promising Results

The tool's headline capability is straightforward: open any photo, tap the hamburger menu, select Tools, then Extend. Users can pinch to zoom out, rotate, and reposition the frame, and the AI fills the newly exposed areas with generated imagery. The feature is also integrated into the Crop tool, making it intuitive for anyone familiar with basic photo editing. In testing, generation times averaged 10-15 seconds per edit, though results varied dramatically based on image complexity.

Key Fact 1: Integration with Crop Tool

One of the smartest design decisions is embedding Extend within the familiar Crop interface. This reduces the learning curve and makes the feature feel like a natural extension of existing workflows. Users can simply drag the corners of the crop box beyond the original image boundaries, and the option to extend appears automatically.

Key Fact 2: Varied Success on Different Subjects

Testing revealed a clear pattern: Extend excels on images with repetitive textures like water surfaces, foliage, or solid backgrounds, but struggles with complex scenes containing multiple distinct objects. For example, a photo of a friend taken with a Nikon mirrorless camera produced nearly seamless expansions that fooled even the subject herself. The AI correctly maintained the shallow depth of field, blurring generated background elements appropriately.

Key Fact 3: Obvious AI Artifacts on Complex Scenes

Conversely, photos with dense details like table settings with multiple plates, glasses, and cutlery often showed telltale AI artifacts. Expanding a mango pastry image produced multiple versions where the color of the tablecloth shifted, the pattern of the background wall warped, and an extra plate appeared on the left that never existed. A second test involving a cafe scene with chairs and tables resulted in blurry, dreamlike textures at the top and oddly shaped bags.

Key Fact 4: Environmental Portraits Shine

The most impressive results came from environmental portraits with natural depth. A silhouette of a tree with plants in the foreground showed expertly generated leaf textures, though the top branches did give away the AI's work. Another photo featuring a person in front of a mirror with reflections demonstrated the tool's ability to maintain realistic blur and lighting consistency.

Key Fact 5: Landscape Expansions Are Outstanding

Landscape photos benefited most from Extend. A picture of a water bed extending toward mountains generated convincing water textures, hills, and a canopy stand. While a car appeared on the right and the subject's arm looked slightly unnatural, the overall result was suitable for social media sharing. A similar test with a gift box on a table produced authentic-looking lamps and wall hangings that never existed in the original shot.

Key Fact 6: Strengths in Simple Compositions

The cleanest results were achieved on images with minimal subject matter—a simple tabletop, an empty room, or a uniform sky. In these cases, the AI's generation was virtually indistinguishable from real photography. This suggests the tool is best used for adding peripheral context to straightforward compositions rather than attempting to rebuild complex scenes.

Key Fact 7: Internet Connection Required

Unlike some on-device AI features, Extend requires an active internet connection and may fail on slow networks. This limitation could frustrate users in areas with poor connectivity, though Apple likely processes the generation on its servers for faster and more consistent results.

Key Fact 8: Metadata Tagging for Transparency

All edited images receive a clear tag in the Photos app's metadata panel, visible when swiping up on the photo. This is a welcome transparency feature that helps users and viewers distinguish between original and AI-enhanced images.

Key Fact 9: Case Examples

Among the test samples, a selfie taken in front of a dam in northern India showed excellent expansion of water and sky, though the foreground rocks exhibited slight texture loss. Indoor shots of food on a table varied: plastic spoons were extended accurately, but the overall scene occasionally lost structural coherence. The sunlit tree and flower photo exhibited the most obvious AI fingerprints, with shrubs, leaves, and petals rendered in a goofy, artificial style.

Key Fact 10: Overall Verdict

Despite its inconsistencies, the Extend tool represents a meaningful addition to Apple's AI lineup. For casual edits where perfection isn't critical—like adding an ultrawide perspective to travel photos—it works well enough that many users will find it genuinely useful. The failures are often subtle, and with time, the technology will only improve. For now, it's a powerful option that expands creative possibilities without requiring professional editing skills.

Source:Digital Trends News