Supreme Court AI Copyright Ruling 2026: AI CANNOT Own Copyright [What Creators Must Know]

Supreme Court AI copyright ruling 2026 coverage needs one important correction. The Supreme Court did not publish a merits opinion saying that artificial intelligence can never be connected to a copyright. It denied a petition for review in Thaler v. Perlmutter, leaving a lower-court decision intact.

The case involved artwork that the applicant identified as autonomously generated by an AI system and listed the machine as the sole author. That fact pattern is different from a person using AI as an assistive tool, selecting material, arranging outputs, writing original text or making creative modifications.

The practical rule for creators is therefore narrower than the headlines. Purely machine-generated material is not protected by copyright under the U.S. position described in the relevant decisions and guidance. Human-authored elements in an AI-assisted work may be protected when they satisfy the ordinary requirements of originality and authorship.

This article explains what the docket entry means, what the D.C. Circuit actually held, what the Copyright Office says about prompts and human contribution and how creators can document their workflow. It is general information, not legal advice.

What You'll Learn

• What the Supreme Court’s certiorari denial means and what it does not decide.
• Why the Thaler facts are different from ordinary AI-assisted creative work.
• How prompts, selection, arrangement and editing can affect the human-authorship analysis.
• What creators should disclose and document when registering work that contains AI material.

What happened on March 2, 2026?

On March 2, 2026, the Supreme Court denied the petition for a writ of certiorari in Stephen Thaler v. Shira Perlmutter, docket No. 25-449. The official Supreme Court docket records the event as “Petition DENIED.”

A certiorari denial means the Court declined to review the case. It does not normally provide a Supreme Court merits opinion, and it does not turn every statement in the lower-court record into a new nationwide rule announced by the Supreme Court. The judgment that was challenged remains in effect for the parties and the legal setting in which it was issued.

The lower-court decision came from the U.S. Court of Appeals for the D.C. Circuit on March 18, 2025. The court affirmed the refusal to register one artwork because the application listed a non-human AI system as the sole author and described the work as autonomously created by a machine.

The timeline is more precise than saying that the Supreme Court “ruled that AI cannot own copyright.” The case concerned authorship of a particular work under a particular application record.

How the case started with an AI-generated artwork

Dr. Stephen Thaler created an AI system called the Creativity Machine. It generated an image titled “A Recent Entrance to Paradise.” In the copyright application, Thaler listed the Creativity Machine as the author and himself as the copyright claimant or owner.

The application also described the image as created autonomously by a machine. That description mattered. The Copyright Office assessed the claim that was submitted. It did not have to decide whether a human who used AI to produce a work could ever own rights in human-authored elements because the application presented the machine as the sole author.

The dispute moved through the Copyright Office, the federal district court and the D.C. Circuit. The D.C. Circuit opinion, hosted in the Copyright Office’s case-document collection, explains that the work was denied because the Copyright Act requires eligible work to be authored in the first instance by a human being.

This is the fact pattern creators should compare with their own workflow. If a person writes the original script, makes creative selections, directs a composition and substantially edits the output, the analysis is not identical to a claim that a machine independently created the whole work.

The site’s vertical AI agents guide makes a similar point about labels and system boundaries. Calling a process AI-generated does not answer what a human actually contributed.

What the D.C. Circuit actually held

The D.C. Circuit held that the Copyright Act requires eligible work to be authored in the first instance by a human being. Because the application named the Creativity Machine as the sole author, the refusal to register that artwork was affirmed.

The court relied on the structure and language of the Copyright Act. Copyright vests initially in an author, terms refer to the author’s life, inheritance, domicile and signature and machines are treated as tools or devices rather than legal authors. The court also discussed the Copyright Office’s longstanding human-authorship practice.

The opinion did not decide every argument a creator might raise after using an AI system. It specifically noted that Thaler’s argument that he should be treated as the author because he created and used the Creativity Machine had been waived before the agency. That issue was therefore not resolved on the merits in the appeal.

The distinction is important. A holding about a work presented as autonomously created by a machine should not be inflated into a rule that the use of an AI tool removes copyright from every resulting work. The Copyright Office’s later Part 2 analysis addresses that broader mixed-work question separately.

What the Supreme Court denial does and does not mean

The denial leaves the D.C. Circuit judgment in place. It also means the Supreme Court did not use this case to issue a new opinion addressing the full range of AI copyright questions. There is no basis for describing the docket entry as a detailed Supreme Court test for prompts, model training, datasets, style imitation or every AI-assisted work.

For U.S. creators, the denial is still significant because it removes the immediate possibility that the D.C. Circuit’s judgment would be reviewed in that case. The lower-court reasoning remains an important reference for applications that claim a machine itself is the author.

It does not mean the Copyright Office has stopped examining mixed human-AI works. It does not mean an AI tool cannot be used in a copyrightable creative process. It does not mean ownership and authorship are identical questions in every work-made-for-hire or licensing arrangement.

Use the official Copyright Office AI initiative page for the source documents. It lists the Part 2 report on copyrightability, registration guidance, review-board decisions and the Thaler court filings.

Can AI-assisted work be copyrighted?

The Copyright Office’s Part 2 report says that AI used as an assistive tool does not by itself remove copyright protection from the output. It also says that purely AI-generated material, or material in which there is insufficient human control over expressive elements, is not protected.

The analysis is case-specific. Copyright can protect original human expression in a larger work that also includes AI-generated material. It can also protect creative human selection, coordination or arrangement of material, and creative modifications that meet the ordinary standard for protection. The AI-generated portions themselves do not become protected merely because they appear next to protected human work.

That means a single file can contain different categories of material. A person’s original article may be protected even if it contains an unprotected machine-generated illustration. A creative layout may be protected as an arrangement even if individual generated elements are not. The final result depends on the facts and the nature of the human contribution.

This is an information framework, not a prediction of how a court or examiner will decide a particular work.

Are prompts enough to create copyright?

The Copyright Office’s Part 2 report says that prompts alone do not provide sufficient human control over expressive elements based on generally available technology. A prompt can communicate a topic, subject, style or desired result, but the system may determine the words, composition, arrangement, rhythm or other expressive details.

A detailed prompt can involve real creative effort. That does not automatically make the resulting output a human-authored work under the current U.S. guidance. The relevant question is whether the person exercised enough control over the expression that appears in the final work.

Iterative prompting, choosing among outputs and editing the result may contribute to the overall human-authored work. The contribution still needs to be evaluated by its creative substance. A high number of prompts is not a legal measurement of authorship.

The site’s AI evaluation guide discusses a related operational problem: automated scores are not the same as proof of origin. A detector cannot replace a record of how the work was made or a legal analysis of human contribution.

What creators and businesses should disclose

The Copyright Office’s 2023 registration guidance says applicants should disclose the inclusion of more-than-de-minimis AI-generated material and briefly describe the human author’s contribution. The applicant should not list the AI system or provider as an author merely because the tool was used in the creative process.

When registering a mixed work, the claimant should identify the human-authored portions and exclude the AI-generated material from the claim where required. The Office’s guidance gives examples involving human-authored text, creative selection and arrangement and human modifications.

Businesses should also preserve internal documentation. Keep the original brief, source material, meaningful revisions, approval records, model or tool used, final human edits and the date of the workflow. Documentation does not guarantee a copyright result, but it makes the factual record easier to explain.

The right disclosure depends on the work and the jurisdiction. When the commercial value or legal risk is material, consult a qualified copyright professional rather than relying on a headline.

Registration and rights-protection workflow

Start by separating the work into human-authored material, AI-generated material and combined elements. Do not assume that the file format or the presence of a human account holder answers the question.

Next, describe the human contribution in plain language. Identify original text, drawings, musical composition, editing, selection, arrangement, sequencing, compositing or other creative work. If the person only provided a prompt and accepted the output without meaningful control over expression, do not overstate that contribution.

Then check the applicable registration instructions. The U.S. Copyright Office says that applicants should use the standard application, identify human authors and disclose AI-generated material. It also explains how applicants may correct a pending or completed record when relevant AI material was not disclosed.

Finally, review contracts. A client agreement should address whether AI tools may be used, who is responsible for third-party rights, what human review is required and whether the delivered work includes unprotected material. Copyright registration is not a substitute for permission to use another person’s work, voice, likeness or confidential information.

The site’s AI design platform guide covers tool capability from a product perspective. For rights management, the workflow record and contract language matter more than the marketing label.

What the ruling means for different groups

For individual creators, the practical lesson is to keep human expression visible and document the process. AI can accelerate ideation, drafting and variation, but a creator should know which portions were generated, which were selected and what was actually rewritten or composed.

For businesses, the issue affects procurement, licensing, marketing claims and internal content policies. A company may own contractual rights in a deliverable while still needing to understand which components are copyrightable and whether third-party material was used.

For developers and platform providers, the case does not create a general license to label generated output as owned content. Product terms, user disclosures, training-data questions, privacy and third-party rights remain separate issues.

For publishers, a human editorial process should be recorded when AI assists with research, drafts, images or translations. An AI disclosure policy can specify review, correction, attribution, source verification and approval requirements.

The site’s agentic AI risk guide is relevant to this governance point. More automated production increases the need for an accountable human process.

Global comparison without overclaiming

U.S. guidance is not a universal global copyright rule. Other jurisdictions may use different statutes, doctrines or registration practices for computer-generated works, related rights and the role of human contribution. A creator should identify where protection will be claimed and where enforcement may occur.

Do not treat a comparison table as a substitute for local advice. Even within one country, the result can depend on the type of work, the human contribution, the application record, contractual terms and the particular court or office reviewing the issue.

Use the Copyright Office’s AI initiative page and the relevant local authority for current information. Avoid stating that the March 2026 docket entry changed copyright law worldwide.

A practical workflow for AI-assisted creators

Before creating, write down the intended human contribution. Decide whether AI will be used for brainstorming, reference, transformation, draft generation, variation, cleanup or another role. Choose the role that leaves a person responsible for the final expressive decisions.

During creation, preserve meaningful versions. Save the source files, major selections, rejected outputs and human edits. If a person combines several generated elements into an original composition, record the selection and arrangement decisions rather than relying only on the final export.

Before publication or registration, review the work for originality, third-party material, privacy, attribution and contractual commitments. Separate the portions that were generated by AI from the portions that a person authored or creatively modified. If the answer is unclear, stop and obtain professional advice.

After publication, maintain a correction path. If the tool, source or disclosure was misidentified, update internal records and assess whether a registration or contract needs correction. Do not wait for a dispute to discover that nobody knows how the work was produced.

The site’s AI workflow analysis and business AI comparison explain why tool capability does not remove the need for an accountable process. Copyright questions are one part of that process, not a setting inside the model.

Bottom line and legal limitations

The March 2, 2026 Supreme Court entry in Thaler v. Perlmutter was a denial of certiorari. It left the D.C. Circuit’s decision in place for an application that identified an AI machine as the sole author of an autonomously generated artwork. It was not a broad Supreme Court merits opinion about every use of generative AI.

The current U.S. framework is more nuanced. Purely AI-generated material is not protected under the Copyright Office’s stated position. Human-authored expression, creative selection, arrangement and modifications may be protected when they meet the ordinary requirements, and the result must be assessed case by case.

The safest operational rule is simple: document what the person created, what the tool generated, what the person selected or changed and what the registration or contract actually claims. Do not rely on a headline, a prompt count or an automated detector as a substitute for the factual record.