As we have written here on this blog, artificial intelligence (“AI”) programs and modules are causing turmoil in the area of copyright law. In general, three areas are concerning to copyright holders:

• What copyrighted inputs are being used by AI programs (for training purposes, for example)?
• What copyrighted content is being permanently held by AI programs as necessary to their functioning?
• What is copyrightable about AI generative output?

The vast amount of pending AI/copyright litigation is focusing on the first two questions and the litigation is now going global. For example, in November 2024, a large Indian news agency, ANI, filed a copyright infringement lawsuit against OpenAI in the Delhi High Court. Recently, legal opinions have been issued stating that the High Court has jurisdiction to hear the case. ANI’s claims are similar to those raised by other news and media organizations: OpenAI is using copyrighted materials without permission or payment to train its AI programs. See here for the media report.

In the meantime, the U.S. Copyright Office has issued a recent report and opinion on the third question — the copyrightability of AI output. The report can be seen here. The report clarifies and reiterates the Copyright Office’s stance on various points related to the copyrightability of AI-generated output. For example, the report reiterates the Office’s rejection of any attempt to make AI-generated output copyrightable. The report states: “Copyright does not extend to purely AI-generated material or material where there is insufficient human control over the expressive elements.” The Copyright Office has said this over and over again. Further, the report also reiterates that human artistic expression is copyrightable even if the work also includes AI-generated material. Essentially, the human authorship is copyrightable, while the separable AI-generated output is not. Again, this position has been often repeated by the Copyright Office. The report also reiterates that copyright decisions involving combined human and AI-generated output are to be made on a case-by-case basis.

Unfortunately, the report also included statements that raise concerns about clarity and about how much time will be spent by the Copyright Office in trying to define and explain a new standard. The central statement in question is this: “Human authors are entitled to copyright in their works of authorship that are perceptible in AI-generated outputs, as well as the creative selection, coordination, or arrangement of material in the outputs, or creative modifications of the outputs.” (emphasis added). The use of the word “perceptible” is obviously worrisome since it is predictable that many people can and will disagree about whether human authorship is “perceptible.”

The report’s other bullet-pointed summary items include the following.

• Questions of copyrightability and AI can be resolved pursuant to existing law without the need for legislative change
• Based on the functioning of current generally available technology, prompts do not alone provide sufficient control.
• The case has not been made for additional copyright or sui generis protection for AI-generated content

Contact the AI, Internet Law and Social Media Attorneys At Revision Legal

For more information, contact the experienced AI, Internet Law and Social Media Lawyers at Revision Legal. You can contact us through the form on this page or call (855) 473-8474.

Understanding the Copyright Office’s “Perceptible Human Authorship” Standard

The Copyright Office’s Part 2 AI Copyrightability Report introduced a standard that has generated substantial debate: human authors are entitled to copyright in works of authorship where human creative expression is “perceptible” in the final output. This framing is a refinement — not a reversal — of the Office’s longstanding requirement for human authorship under Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884), and the more recent Thaler v. Perlmutter, 654 F. Supp. 3d 76 (D.D.C. 2023), which flatly rejected registration for a work created autonomously by an AI system without any human author.

The practical question is what makes human expression “perceptible” when a human uses an AI tool. The Copyright Office has indicated that the answer depends on the degree of human creative control exercised over the expressive elements of the final work. A human who writes a detailed narrative framework, specifies stylistic choices, curates generated outputs, selects specific AI-generated elements and discards others, and then arranges and integrates those elements into a coherent whole is exercising the kind of creative authorship that can support copyright registration — for the human-authored portions.

What the Report Actually Allows: Registrable Human Authorship in AI-Assisted Works

The Copyright Office has now confirmed several categories of human creative contribution that are registrable even when AI tools are involved in the creative process:

• Selection and arrangement — A human who selects which AI-generated images, text passages, or musical phrases to include in a compilation, and who arranges them in a creative order, can claim copyright in that selection and arrangement. This mirrors the longstanding rule for compilations under 17 U.S.C. § 103.
• Original text integrated with AI output — Where a human author writes original text and incorporates AI-generated text into the work, the human-authored text is protectable. Registration is available for the human-authored portions, with the AI-generated portions disclosed as unprotected.
• Modifications and creative edits — Where a human significantly modifies, edits, or transforms AI-generated content through creative choices, those modifications can constitute registrable authorship.
• Visual works with human-authored elements — In graphic novels like Zarya of the Dawn (the subject of a 2023 Copyright Office registration decision), the human author’s written text and the arrangement of the work were registrable, while the AI-generated images were not.

The Disclosure Obligation: Registering AI-Assisted Works Correctly

When filing a copyright registration application for a work that includes AI-generated material, the applicant must disclose the use of AI and limit the claim to the human-authored elements. The Copyright Office’s updated guidance on AI-assisted works requires applicants to:

• Exclude AI-generated content from the copyright claim by describing what the human author created
• Provide a statement identifying the AI-generated portions of the work that are being disclaimed
• Describe the human author’s specific creative contributions in the application

Failure to disclose AI involvement — and then claiming copyright over AI-generated material — exposes the registration to invalidation and potentially to fraud-on-the-Copyright-Office challenges. The Office has made clear it is actively reviewing AI disclosures as part of the examination process.

The Training Data Litigation: A Parallel and Unresolved Front

While the copyrightability of AI output is becoming clearer, the training data litigation involving AI companies remains in its early stages. In the United States, the major pending cases include Getty Images (US), Inc. v. Stability AI, Ltd. (D. Del.), Andersen v. Stability AI (N.D. Cal.), and cases brought by The New York Times, music publishers, and other content owners against OpenAI and Microsoft. These cases raise the fundamental question of whether training a large language model or image generator on copyrighted works without a license constitutes copyright infringement — or whether such training qualifies as fair use under 17 U.S.C. § 107.

Courts have not yet issued definitive rulings on the merits of these cases. The stakes are enormous: if training on copyrighted content without a license is held to be infringement, the business models of virtually every major AI company would require restructuring. Fair use analysis in this context will likely require courts to weigh the transformative nature of AI training, the market impact on copyright owners, and whether the AI outputs substitute for the original works in ways that harm rights holders.

Practical Implications for Businesses and Creators

For businesses and content creators using AI tools, the current legal landscape creates both opportunities and risks:

• Register your human authorship now. If you are producing works with significant human creative input alongside AI tools, register the human-authored elements. Copyright registration remains a prerequisite for filing an infringement suit in federal court under 17 U.S.C. § 411.
• Document your creative process. Keep records of your creative decisions: what prompts you used, what outputs you rejected, what edits you made, and how you arranged the final work. This documentation will be critical if the registrability of your work is ever challenged.
• Review your AI tool agreements carefully. Many AI platforms claim broad licenses over outputs generated using their tools. Review terms of service to understand what rights the platform retains and whether you can claim ownership of the work product.
• Watch the training data litigation. A ruling that training on copyrighted content constitutes infringement could significantly affect the availability and cost of AI tools. Businesses that rely heavily on AI-generated content should monitor these cases closely.

Contact the Copyright and AI Attorneys at Revision Legal

The law governing AI and copyright is developing rapidly. The experienced Copyright Attorneys at Revision Legal help creators, businesses, and technology companies navigate copyright registration, licensing, and infringement issues involving AI-generated and AI-assisted content. Contact us through the form on this page or call (855) 473-8474.