Apple Embraces AI Photo Editing With Google Watermarks
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Apple announced at WWDC 2026 a suite of AI-powered photo editing tools, including prompt-based manipulation and an updated Image Playground that introduces photorealistic image generation for the first time, a shift from its prior cartoon-only approach.
- Clean Up, Extend, and Spatial Reframing — Apple's new trio of Photos app tools — will embed Google-developed SynthID watermarks to flag AI-manipulated images, and the updated Clean Up now offers more realistic infill even in complex scenes.
- Extend can expand an image beyond its original dimensions using generative fill, similar to Adobe's Generative Expand in Photoshop, while Spatial Reframing lets users drag parts of a photo to simulate a different shooting angle, building on Vision Pro's spatial modeling.
- Craig Federighi said two years ago that Apple wanted to "purvey accurate information, not fantasy," and when Apple Intelligence launched in 2024 he said the company was "concerned" about AI undermining trust in photographic content as a record of reality.
- SynthID adoption is expanding industry-wide — OpenAI recently adopted it — and users can check images for the watermark by uploading them to Gemini or Google's AI-powered Search, though the article notes most synthetically generated imagery online remains unlabeled.
- Image Playground in Apple's demo generated a woman holding a birthday cake using a real photograph as reference, entirely replacing the original background — the kind of photorealistic manipulation Apple had previously avoided to prevent believable deepfakes.
Why it matters: Apple is betting that SynthID watermarks — which the article notes are still in early adoption and leave most manipulated imagery online unlabeled — are sufficient to let it chase Google's and Samsung's generative editing features without fueling misinformation. The reversal from Federighi's 2024 statement that Apple was "concerned" about AI distorting reality makes this a credibility test: if watermarks fail to scale across platforms, Apple will have helped normalize photorealistic manipulation of personal photos.
