AI & ML Practical Magic

A new tool reverses the generative AI enhancements inside modern cameras to reveal the raw, unedited reality of a photo.

April 25, 2026

Original Paper

Addressing Image Authenticity When Cameras Use Generative AI

Umar Masud, Abhijith Punnappurath, Luxi Zhao, David B. Lindell, Michael S. Brown

arXiv · 2604.21879

The Takeaway

Smartphone cameras now use AI to fill in details that the lens cannot actually see. This process creates hallucinated pixels that look real but are essentially fabrications. A new recovery method allows users to strip away these AI additions to see the original sensor data. We are losing the ability to trust photos as historical records as AI becomes part of the image signal processor. This technology provides a way to verify what was actually present in front of the lens.

From the abstract

The ability of generative AI (GenAI) methods to photorealistically alter camera images has raised awareness about the authenticity of images shared online. Interestingly, images captured directly by our cameras are considered authentic and faithful. However, with the increasing integration of deep-learning modules into cameras' capture-time hardware -- namely, the image signal processor (ISP) -- there is now a potential for hallucinated content in images directly output by our cameras. Hallucina