On 2 March 2026 the US Supreme Court issued its order list, denying Dr Stephen Thaler’s petition for certiorari in Thaler v. Perlmutter (No. 25-449). The petition sought to challenge the position that works generated entirely by artificial intelligence, without human creative contribution, cannot be registered for copyright. By declining to hear the case, the Court leaves the D.C. Circuit’s earlier decision in force nationwide.

What the case was about

Thaler argued that an image autonomously generated by his AI system, DABUS, should be registrable for copyright. The US Copyright Office refused, on the basis that copyright law requires a human author. Thaler litigated up to the appellate court, lost, and petitioned the Supreme Court.

The denial does not mean that all AI-generated content is unprotectable. It confirms a narrower rule: purely machine output, with no human creative contribution at all, is not a protected work.

Why this matters to Chinese companies too

First, it draws the line between “AI-assisted creation” and “autonomous AI generation”. Where there is substantial human creative contribution, protection may still be available; where there is none, it is not. Companies seeking registration or asserting rights must be able to evidence human contribution — prompts alone are usually not enough, as we discuss in our note on why a prompt list is not a work.

Second, it affects the asset structure of cross-border content businesses. A content library consisting mainly of AI-generated material may not be bookable or pledgeable as a copyright asset in the US market, which feeds through to financing and M&A valuation.

Third, it does not conflict with criminal exposure. The absence of copyright protection does not mean others’ works can be used freely; using AI as a tool to copy and distribute others’ original artwork for profit may still constitute a copyright crime — see our note on the Beijing High Court typical case.

Three practical steps for AI companies

First, keep evidence of the creative process. The full chain — prompt iterations, human selection, post-editing and final approval — is the basis for asserting rights. For a fuller approach, see our IP protection service.

Second, distinguish “outputs” from “products”. A company’s real assets are often the model, the toolchain and the workflow rather than individual images, and their ownership paths differ from content copyright.

Third, assess title before going global. The US and the EU differ on the protectability of AI-generated content; on the EU side see our EU AI Act service.

The significance of Thaler is not that it changed anything, but that the Supreme Court’s silence confirms that human creative contribution remains the floor of the copyright system. For companies whose core assets are AI-generated, that floor belongs on page one of product and asset design.

This article is based on the US Supreme Court’s 2 March 2026 order list and public reporting. It is for general reference only and does not constitute legal advice.