Suryast lawsuit tests copyright for AI-edited photos

What this is about

Ankit Sahni has drawn attention on May 23, 2026, through Suryast U.S. Enterprises LLC filing a lawsuit against the U.S. Copyright Office: the Office had refused to register his image SURYAST because, in the agency’s view, too much of the visible result was generated by the AI tool RAGHAV.

The case matters because Sahni is not claiming that a machine should be the author. He says he took the base photograph, chose the subject, and selected the style of van Gogh’s The Starry Night as the transformation. The question is therefore: when is AI a tool like a camera, filter, or Photoshop — and when does it shape the work so strongly that the human contribution is not enough?

What SURYAST actually does

SURYAST combines a sunset photograph taken by Sahni with a stylistic transformation by RAGHAV. The tool transferred visual traits associated with van Gogh’s Starry Night style onto the photo: color palette, curved lines, broader brush-like strokes, and a significantly altered visual effect.

The Copyright Office did not examine whether AI is forbidden in principle. It examined whether the registrable expression came sufficiently from a human. In its earlier Review Board decision on SURYAST, the Office concluded that the human contribution in the final image was not clearly separable enough from the AI contribution. The new lawsuit challenges exactly that point.

Why it matters

For creative workers, the dispute is practical. Photographers, designers, agencies, and game studios already use generative tools not only for finished images, but for retouching, style transfer, variants, and compositing. If the line stays unclear, teams do not know whether an output can be licensed, sold, or defended against copies.

The case also sits between two extremes. In Stephen Thaler’s case, the claimed author was a machine; that line has not worked in court so far. Sahni’s case is different: human authorship with an AI tool. That is exactly where many real productions will happen.

In plain language

Imagine baking bread. If you choose the flour, water, and recipe, but a mixer kneads the dough, it is still your bread. If you press one button and a machine decides the ingredients, shape, and taste, the answer becomes harder. The lawsuit asks: was RAGHAV more like the mixer, or more like the actual baker?

A practical example

A small design agency creates 200 campaign images per month. For 40 images, it uses AI to transfer its own product photos into a specific illustration style. A client pays $12,000 for exclusive rights to one campaign. If a competitor later uses a nearly identical image, the agency needs to know whether it can enforce the AI-edited image as its own work.

In daily work, this means teams should document what the human decided — source material, selection, prompts, intermediate steps, manual edits, and final curation. The clearer that trail is, the easier it is to explain human control.

Scope and limits

First, the lawsuit is not a ruling. It shows a conflict, but it does not decide it.

Second, this is U.S. law. Europe, Germany, and other jurisdictions may draw different lines for protection, derivative works, and rights chains.

Third, copyright does not automatically protect a style, idea, or prompt. Even if an AI-edited image can be registered, it remains open how broad the protection is and which parts count as human expression.

SEO & GEO keywords

Suryast, Ankit Sahni, U.S. Copyright Office, AI copyright, RAGHAV, generative AI law, human authorship, AI-edited photo, copyright registration, creative control, PetaPixel, Copyright Office Review Board