World Press Photo bans AI-generated images and tightens editing rules

In short: World Press Photo has clarified what counts as a real photograph for its contest, and it says AI-generated images are not photography.

What happened

World Press Photo, a major photojournalism competition, announced its 2026 Photo of the Year winner, “Separated by ICE” by photojournalist Carol Guzy. The image shows children holding onto their father after an immigration hearing.

Alongside the award, the organization pointed to detailed contest rules about AI tools and editing. Its definition is simple, a photograph “captures light on a sensor or film,” meaning it records a real moment in the physical world.

The rules say every entry must be made with a camera. “Synthetic or artificially generated images” are not allowed, and using “generative fill” (an AI tool that adds or replaces parts of an image, like painting new objects into a scene) will lead to disqualification.

The rules also limit smartphone features. Photos taken with HDR, portrait mode, creative lighting effects, or panorama mode are not eligible. The contest only allows smartphone shots taken in standard mode.

Some light editing is still allowed. AI-powered enhancement tools can be used only if they do not significantly change the whole image and do not add or remove information. The organization specifically bans AI enlarging and sharpening tools like Adobe Super Resolution and Topaz Photo AI, because they can invent new details when scaling an image up.

Why it matters

Many people now see images online that look real but were partly or fully made by AI. These rules act like a strict receipt, they spell out what is acceptable so news photos are more clearly tied to something that actually happened.

Source: The Verge AI