Who Owns an AI Generated Display Render?

The question usually comes up at the worst moment. The concept is approved internally, the client wants the file for their own deck, and someone at the manufacturer asks, half joking, whether we're actually allowed to hand over an image a machine drew. Then procurement sends a supplier agreement saying every deliverable is original work that transfers in full.

I've worked in point-of-purchase for close to four years and I build AI POP Displays. The short answer is that the tool's terms leave the render with you, and that's the easy layer. The harder layer is whether there's any copyright in it to hand over, and in the US that depends on how much of the render a person made.

This is what to check and what to ask for, not legal advice. Your lawyer writes the clause.

Who owns an AI render the moment it's generated?

Three layers decide it, and people mix them up constantly.

The tool's terms say who holds whatever rights exist between you and the vendor. Copyright law says whether any rights exist in the first place. Your contract with the client says who ends up with them. A render can be yours under the first, partly unprotected under the second, and still promised in full to a client under the third. That combination causes the trouble.

On the first layer the vendors I checked land in the same place, with different conditions.

Google. The Gemini API terms say Google won't claim ownership of the content it generates for you. They also say Google may generate the same or similar content for other users. AI POP Displays runs on Google's Nano Banana Pro, through that API on paid terms.

Midjourney. Its terms say you own the assets you create to the fullest extent the law allows, subject to the agreement and to third parties' rights. A company with more than $1 million a year in revenue, or an employee of one, has to be on the Pro or Mega plan to own them. Content is publicly viewable by default unless you're on a plan with Stealth mode.

Two details matter for display work. First, "you own it" in a vendor's terms means the vendor isn't claiming it. It doesn't create a copyright the law doesn't recognize. Second, the output isn't exclusive. Google's terms say so outright. A generic prompt for a "premium acrylic counter display" can produce something close to what another user gets, which is one more reason a concept should start from the client's real product and a sketch rather than from a phrase.

Can an AI render be copyrighted in the US?

Partly, and only the part a person contributed.

The US Copyright Office published its report on copyrightability and AI in January 2025. Its position is that copyright doesn't extend to purely AI-generated material, and that prompts alone don't give a user enough control over the output to be its author. What can be protected is human work that's perceptible in the output, the creative selection, coordination or arrangement of material, and creative modifications of the output. It's decided case by case.

Two decisions show the line. In the Zarya of the Dawn letter of February 2023, the Office registered the author's text and her selection and arrangement of the comic's elements, and excluded the Midjourney images themselves. In Thaler v. Perlmutter, the D.C. Circuit held in March 2025 that the Copyright Act requires a human author, while noting that the rule doesn't bar works made with the assistance of AI. The Supreme Court declined to take the case in 2026, so that ruling stands.

The Office's March 2023 registration guidance adds a practical duty. If you register a work that contains AI-generated material beyond a trivial amount, you disclose it and exclude it from the claim.

For a display concept, this sorts the inputs neatly. A designer's own sketch that the render follows is human work, and it stays visible in the result. The choice of which variant goes to the client, and what gets changed in each round, is human selection and modification. The surface the model filled in on its own (the reflections, the shop lighting, the texture of the acrylic) is the part the Office says isn't protected. The more of the concept that came from your sketch and your decisions, the more of it is yours in the legal sense too.

Whose artwork is inside the render?

This is the question nobody in the copyright articles asks, and on a display it's the first one. A POP concept almost never shows an invented product. It shows the client's pack, the client's logo and the client's key visual on a header card, because that's what the client is approving.

None of that changes hands when it lands in a render. The logo, the packaging design and the product photography belong to the brand, or to whoever the brand licensed them from. The render is a composite, and the brand's artwork sitting inside it remains the brand's. Your rights, whatever they are, cover what you added around it.

Two consequences follow. You need the client's permission to put their artwork into a tool at all, which is a confidentiality question as much as an ownership one, and I went through what each kind of tool does with an upload in can I put my client's artwork into an AI display tool. And you shouldn't put a third party's artwork in without permission either. A competitor's pack used "for scale" or a stock photo pulled from a search can end up in a deliverable you've promised is clear.

This is also where the way a tool takes references starts to matter. If the render is built from the client's real pack shots, labeled as product and placed by role, the client's artwork is in the concept on purpose and in the right place, rather than something the model half invented from a description. If you're about to build a concept from a client's own packs, AI POP Displays is free to start and reads each reference by its role, with your sketch last.

Does owning the render mean owning the display design?

No, and for a manufacturer this is the distinction worth keeping straight. A render is an image. Copyright in an image, whatever its extent, doesn't protect the physical display somebody builds from it.

The physical design has its own protection in the US. Under 35 U.S.C. 171, a "new, original and ornamental design for an article of manufacture" can be protected with a design patent, which lasts 15 years from grant. In November 2025 the USPTO rescinded its 2024 guidance on AI-assisted inventions and replaced it with revised guidance that applies one standard whether or not AI was used. AI systems are tools, only natural persons can be inventors, and the notice says explicitly that this covers design patents as well as utility patents.

Most POP displays are never patented, but the principle carries over. The design of a display lives in decisions a person makes: the format, the proportions against the product, the material, how many facings, where the brand sits. A render records the look of those decisions. It doesn't carry the dimensions, the material per part or the joints, which is why a manufacturer can't quote from it, as I explained in how a brand gets an AI concept built.

If the design is the client's, the useful record is the one that shows who made those decisions and when. The brief, the sketch, the variant chosen and the edit asked for each round are that record.

What do AI POP Displays' terms say about the renders?

You keep all rights in the briefs, reference images and product images you upload, and you grant us a license only to process them to deliver the service. For the renders, we assign to you all rights we may have in them, to the extent they're assignable.

Two other lines are there because of everything above. You're responsible for making sure your use of a render doesn't infringe someone else's rights, meaning trademarks, packaging, designs or photographs you reference. And we don't warrant that a generated render is itself protected by copyright, because the law varies by jurisdiction and, as the Copyright Office report shows, depends on the human contribution.

Nothing you upload or render trains AI models, on any plan, including the free one. For a concept that carries a client's unreleased launch, that's the clause that lets you say yes when the brand asks where the artwork went. If a client is waiting on that answer before you send the next concept, start it on the free plan, which carries the same terms as Pro.

What should go into the client contract?

The supplier agreement was probably drafted before anyone delivered an AI concept. Read its IP clause with these points in mind, and raise them with your lawyer.

1. Say the concept is AI-assisted. Name the tool and the account type. Many brands now have an AI policy you'll need to fit.
2. Assign what you have, on payment. Transfer your rights in the deliverable to the client when it's paid, and avoid a warranty that the AI output itself is protected by copyright. Nobody can promise that today.
3. Get the client's warranty on their own material. The client confirms it owns or has licensed the logos, packaging, photography and key visuals it sends you to use in the concept.
4. Call the render a concept. It shows a look for approval. It isn't an engineering drawing or a manufacturing specification, and the quote, sample and production drawings follow separately.
5. Keep the human record. Save the brief, your sketch, the variants generated, which one was chosen and what was changed in each round, with dates. If ownership or registration ever comes up, that's the evidence of what a person contributed.
6. Keep the tool's terms, dated. Save the version of the terms in force when you made the concept.

Where to go from here

The tool's terms leave the render with you. The law protects the part of it a person made. The client's artwork inside it stays the client's. And the display itself, the physical piece, is a design someone decides, not an image someone generates.

That last point is where we're taking the work next. Bellto is an AI agent for brands and agencies that takes a POP campaign from brief to shop drawings. You tell it what you're launching (brand, product, channel, stores, budget, date) and it asks for what's missing, proposes the campaign mix and materials with the budget split per store, and designs every piece with you at real scale in parametric 3D, with proportions derived from the product and the facings. Every approved piece comes out as a 3D model, a part-by-part cutlist, dimensioned drawings, a STEP of the assembly and a cut DXF per part, a package a workshop can quote without redrawing, and it suggests manufacturers that fit by material, format and volume. We're building it now. The waitlist is open to any brand, we're contacting the first ones soon to run the first real campaigns end to end, and it opens by invitation, in small groups, with the campaign you describe when you sign up setting your place. Pricing goes first to the people on the list, and there's no card.

For the concept your client is asking about this week, AI POP Displays is free to start with 15 one-time credits and no card required, then Pro Beta is $49 a month for 150 concept generations at the founding price, with the $69 list price stated openly. Renders are assigned to you on both. Make the next concept from your own sketch, and if the campaign behind it has to reach a workshop, put it on the Bellto list.