Rights
The work we make for you is yours. This page says what that means, where the limits are, and what we record so your ownership holds up.
Last reviewed 2026-10-14
The work is yours
When you approve a deliverable, we assign to you all the rights we hold in it: the copy, the images, the pages and the emails, for any use, in any medium, with no time limit. Each asset's rights record says assigned to: client, and your workspace shows the rights next to every approved asset.
We do not reuse your deliverables for other clients, and we do not show them in our own marketing unless you agree in writing. Your brand profile, your claims and the material you send us stay yours; we use them only to make your work.
Model terms
The drafts come from commercial AI models called through their providers' APIs, under terms that allow commercial use of what the models produce. The providers do not claim ownership of those outputs.
Every deliverable's AI manifest names the models used for it, and its rights record names the provider terms that applied, for example the Google Gemini and Imagen generative media terms for an ad image.
No training on your data
We do not train or fine-tune any model on your brand profile, briefs, claims, deliverables, feedback or performance data.
Your brand profile is loaded only into the runs we make for you. One client's material is never used to make another client's work.
Indemnity
We do not give an intellectual-property indemnity for AI-generated work by default. Each deliverable's rights record says so: indemnity: none.
What we do instead: every concept is checked against your approved claims, your brand profile and the ad platform's policies, and every image for text, logos, artifacts and people; a person at Plunge reviews your first pieces of each kind, anything with people, a new claim, a regulated product, video or ad spend, and a sample of the rest; and every piece is approved before it is published. Those checks lower the risk; they are not a guarantee. If your company needs an indemnity, raise it before you sign so it can be discussed in writing.
The human-contribution log
Thaler v. Perlmutter (US Court of Appeals for the D.C. Circuit, 2025): copyright in the United States requires a human author, so a work made by an AI system on its own cannot be registered.
That is why the human part of each deliverable matters. Purely machine-generated material may have little or no copyright protection, while the choices and changes people make can be protected. So every deliverable's manifest keeps a log of the human contributions, in order and with a time:
- a steer: who asked for a change to a concept, and what they asked for;
- an edit: who changed the copy, the page text or the image, which parts, when and for what reason (a text edit made in place is a new revision, checked against your approved claims and brand rules before it is saved, and listed in the edit log under the piece; a lockup adds your logo and call to action to an image);
- an approval: who approved it — a person at Plunge, you in your workspace or your approver from an email link.
If you ever need to register or defend a piece, that log is your record of who did what. We cannot promise that any particular deliverable is protected by copyright; the log is the evidence, not the outcome.
What AI may make
Your approver decides, in your workspace under Trust & controls, three things: no AI-generated people, product images only from photos you supplied, and no video. Production and the quality checks follow those switches, and a request for a format you switched off is refused.
A brief that shows a real person's face or uses their voice is refused until your approver records that person's consent; revoking it stops new work that relies on it.
No testimonials, reviews, star ratings or endorsements appear in the work unless they are quoted from claims you approved (the FTC rule on fake reviews and testimonials, 16 CFR Part 465), and the work never sells itself as made by AI.
Questions
Ask us, or write to legal@plungeai.com. See also AI disclosure for how AI-made work is labelled and Privacy for what we keep.