Generative AI in consumer photo editing

Generative artificial intelligence is being integrated into consumer photography and editing tools to simplify complex image manipulation. Early implementations, such as OpenAI’s ChatGPT Images and Samsung’s Galaxy AI, allow users to modify backgrounds, reposition objects, or remove distractions using natural language or native features like ‘AI Eraser’.

As these technologies advance, they have moved toward sophisticated image reconstruction. Tools like Aiarty Image Enhancer can now analyze images to reconstruct missing details and reduce noise. However, this increased capability has introduced new issues. Testing of Apple’s iOS 27 features, including ‘Expand’ and ‘Spatial Reorientation’, revealed that the AI can generate entirely fictional people and faces when attempting to fill in obscured or cropped areas, sometimes inventing strangers based on minimal visual cues.

In response to the rise of deepfakes and AI-generated imagery, Apple is introducing a new photo authentication system called ‘Apple Reference Image’ for the iPhone 18 Pro and iPhone 18 Pro Max. This technology uses a cryptographic signature applied at the hardware level. When a photo is captured in ‘Reference mode’, the camera sensor signs the pixel data immediately to prevent software tampering. The verification process utilizes the device’s Secure Enclave and encrypted timestamps, with final validation performed via Apple’s Private Cloud Compute. To ensure long-term security, the system also applies an additional signature designed to resist future quantum computing attacks.

Further developments in iOS 27.2 include a data-wiping feature that allows users to overwrite storage with empty data for permanent deletion. Due to regulatory requirements in China, this specific deep-cleaning function is expected to be available only for models sold in that region.