Can AI-generated designs be copyrighted?

If you generated an image with AI and assumed it's legally yours, there's a good chance it isn't. Under current U.S. law, works produced solely by artificial intelligence cannot be copyrighted at all — no matter how much you paid for the tool or how detailed your prompt was.

The reason comes down to a single principle: copyright protection requires human authorship. The U.S. Copyright Office has been consistent on this point, stating that it "will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author." That position was formalized in the Office's March 2023 registration guidance and has survived legal challenge — in Thaler v. Perlmutter, a court upheld the Office's refusal to register a purely AI-generated work.

The January 29, 2025 report from the Copyright Office reinforced the stance. Its conclusion: entering prompts into an AI system — even detailed, carefully crafted ones — does not constitute authorship. Prompts express ideas, but they don't control how those ideas are expressed. As one legal analysis of the report puts it, someone specifying what an AI output should look like is no more an author than a person who merely describes a commissioned work to an artist.

For small business owners, this creates a real problem. Many websites today are filled with AI-generated visuals — hero images, featured graphics, illustrations. If those images were produced entirely by AI, no one owns them under copyright. Not you, not the AI company. They effectively sit in the public domain, which means competitors could legally copy them.

There is a workable middle ground. The Copyright Office has granted hundreds of registrations for works that combine human and AI elements since its 2023 guidance — but only the human-authored portions receive protection. Key points to understand:

• Purely AI-generated images cannot be registered or protected by copyright.
• Substantial human modification of AI output can make the human contributions copyrightable.
• Applicants must disclose AI-generated material in registration applications and disclaim those portions.
• Using AI as a tool — for ideation, editing, or refining human-created work — doesn't disqualify the result.

This is why the media creation process at AI Business Sites is built around human creative direction rather than raw AI output alone — because a website's visuals should actually belong to the business that paid for them. The same principle applies to any business commissioning design work: if no human expression is embedded in the final image, no copyright exists to protect it. And the law isn't changing soon — the Copyright Office maintains that existing copyright law is sufficient to handle AI-generated works and has opposed creating new, special legislation for them.

The takeaway for business owners is simple: before you build a website around AI-generated designs, understand what you actually own — and what you don't.

You can spend an hour crafting the perfect prompt — and still end up with an image you don't legally own. That's the uncomfortable conclusion the U.S. Copyright Office reached in its January 2025 report on generative AI outputs, and it catches most business owners off guard.

The Office's reasoning is blunt: "the mere selection of prompts, even if those prompts are detailed and are the product of some human effort, does not itself yield a copyrightable work." Prompts describe what you want — but copyright protects what you actually express, and the AI makes the expressive choices, not you.

The Copyright Office offered a memorable analogy: just as a person who describes to an author what a commissioned painting should look like is not a joint author of that painting, someone entering a prompt specifying what an AI output should look like cannot be considered an author of the result. Commissioning is not creating.

This isn't a fringe position. The Office's official registration guidance states it "will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author." And internationally, countries that have examined the issue "so far have agreed that copyright requires human authorship."

There's a practical middle path. When a work mixes human and AI-generated elements, only the human-authored portions are potentially copyrightable — each assessed case by case for originality and creativity.

The landmark illustration is the Midjourney graphic novel decision. The Copyright Office concluded the work — human-written text paired with Midjourney-generated images — was copyrightable as a whole, but the individual AI-generated images received no protection. The author's words were hers; the pictures belonged to no one.

For businesses using AI in design workflows, the guidance is clear:

• Substantially modify AI outputs so human expression is perceptible and original
• Keep documentation of human creative contribution — sketches, notes, edit logs
• Disclose AI-generated material in copyright applications, or disclaim those portions
• Use AI tools trained on licensed or public domain data to reduce infringement risk

Since the Office's March 2023 guidance, hundreds of registrations for works incorporating some AI-generated content have been granted — proof that a hybrid approach works when handled correctly.

This is exactly why AI Business Sites treats media creation as a managed, human-reviewed process rather than raw generation: every image on a client's site passes through quality checks, human oversight, and brand styling, and the client owns everything the system produces for them — with the compliance question already answered before launch.

To ensure AI-assisted designs receive copyright protection, businesses must treat AI as a tool that augments—not replaces—human creativity. The U.S. Copyright Office has made clear that while purely AI-generated works lack the human authorship required for protection, works where human expression remains perceptible in the final output are eligible for registration. This distinction hinges on demonstrating that the human contribution goes beyond mere prompting and involves substantive creative control over the design’s expression.

Since the Copyright Office’s March 2023 guidance on works containing AI-generated material, hundreds of hybrid registrations have been granted, affirming that AI-assisted designs can be protected when properly documented and structured. For businesses using platforms like AI Business Sites—which generates AI-powered featured images and content as part of its done-for-you website service—this means taking deliberate steps to secure rights in the final deliverables. Simply accepting raw AI outputs offers no copyright protection; instead, substantial human modification is essential to transform the work into an original expression eligible for protection.

To build a defensible copyright claim, creators should follow actionable recommendations grounded in Copyright Office guidance: substantially modify AI-generated designs to reflect personal artistic judgment; document the human creative process through sketches, notes, or modification logs; and when applying for registration, disclose the use of AI and disclaim any non-human-authored portions. These practices not only align with legal standards but also reinforce transparency with clients about the role of AI in the creative workflow—a growing expectation in professional design services.

By integrating these steps into their workflow, businesses using AI-assisted design tools can confidently assert ownership over their visual content while remaining compliant with evolving copyright norms. This approach protects both the creator’s investment and the integrity of the work in commercial use.

The images your AI tool spits out may look original, but if the model learned its craft from millions of scraped copyrighted works, your "original" design could carry someone else's fingerprints. That risk starts at the training data, not the output.

Generative AI systems are trained on millions of copyrighted images, and that creates fundamental tension with existing copyright law, as legal scholars have noted. No court has yet applied the fair use exception to generative AI art programs specifically, which means the legal ground remains unsettled for businesses using these tools commercially.

The U.S. Copyright Office has taken a nuanced position on training data itself. As the Office stated in its AI report, "some uses of copyrighted works for generative AI training will qualify as fair use, and some will not." In other words, you cannot assume the tool you're using is in the clear.

The practical fix is choosing tools with clean provenance. Legal experts recommend selecting AI platforms trained on licensed or public domain datasets — Adobe Firefly, Getty, and Shutterstock are the leading examples — rather than models trained on scraped works without creator consent, per IP law guidance.

To reduce infringement risk when using AI-generated visuals commercially:

• Choose AI tools trained on licensed or public domain datasets rather than scraped copyrighted content
• Apply substantial human modification to AI outputs so original human expression remains perceptible
• Keep documentation of your creative contribution — sketches, notes, and edit logs support any copyright claim
• Disclose AI-generated material in copyright applications, since only human-authored portions receive protection

Transparency matters on the client side, too. If AI tools play a role in a commercial project, experts advise informing clients openly to manage expectations and reduce legal exposure from undisclosed AI use. A business that hides its AI involvement risks more than trust — it risks disputes it cannot win.

This is why provenance should be a procurement question, not an afterthought. When AI Business Sites generates featured images for client websites, the question of where the model's training data came from is treated as part of compliance, not a footnote. Knowing what your tools were trained on is now as important as knowing what they produce, because the Copyright Office processes roughly 500,000 registration applications annually — and undisclosed AI material can jeopardize any of them.

Ask your provider two questions before you buy: what data trained the model, and will you tell my clients AI was involved? Honest answers to both are the foundation of compliant, defensible design work.

Knowing the rules is one thing; building a content pipeline that actually follows them is another. The good news is that the Copyright Office's own guidance makes the compliant path clear — it just takes a workflow designed around human authorship.

Start with the core principle: only human-authored portions of a work are copyrightable, so your process needs to demonstrate real human creative input, not just prompts. As Skadden's analysis of the Copyright Office's January 2025 report notes, using AI as a tool to enhance human creativity — for ideation, editing, or modifying inputs — doesn't disqualify a work from protection, provided human expression remains perceptible and original.

That translates into four practical practices for any business publishing AI visuals:

• Human review and modification — every AI-generated image gets human oversight and creative adjustment before publishing, never used raw from the model.
• Style guides as human direction — documented brand colors, reference imagery, and layout choices represent genuine human creative decisions that shape the output.
• Documented workflows — records showing who reviewed what, when, and what changed, which the Copyright Office's registration guidance supports when claiming human contribution.
• Honest disclosure — applicants must disclose AI-generated material in copyright applications, and legal experts recommend transparency with clients about where AI is involved.

This is exactly how AI Business Sites approaches featured images. Every page — blog posts, service pages, location pages, comparisons — publishes with a 1200×630 featured image generated from a per-client style-guide library: reference images you choose, brand colors injected automatically, and a designed headline baked in. A vision-QA judge scores each image for text legibility and brand fit, auto-approving only high scorers and routing the rest to human review.

The result is a pipeline where the AI does the production work but humans govern the creative direction, the quality bar, and the final call. Since the Copyright Office has granted hundreds of registrations for works incorporating AI-generated content since its March 2023 guidance, a human-governed process is a genuinely defensible one — and it's managed for you, end to end.

Want to see what a compliant, human-governed media pipeline looks like on your own site? Book a free 30-minute strategy call — no obligation, no sales pitch, just a real conversation about your website and goals.