How we use data for AI

Exactly what happens to your briefs, brand assets and ads inside Ori: what the model sees, what we keep, what we never train on, and the controls you have.

Last updated 16 October 2026

The short version

Your content is used to make your ads, and for nothing else, unless you opt in in writing.
Origin isn’t trained on customer briefs, assets, ads or likenesses by default.
Brand memory is per workspace: never shared, always visible, and you can make Ori forget anything.
Close your workspace and everything in it is deleted within 30 days, and from backups within a further 35.

1. What happens to a brief#

You write or say a brief. It’s stored in your workspace, in your data region.
Ori reads it back as a plan. The plan uses your brief, your brand rules and your brand memory. Nothing from other customers is used.
Safety checks run. The brief, script, frames and voice track are checked against the Usage Policy. Anything flagged is held for a trained reviewer.
Origin renders the cut. Your media is processed on GPU servers in your region, and the result is saved to your workspace with content credentials.
You approve, change or discard it. Discarded plans and drafts are deleted with your workspace, or when you delete them.

2. Training#

Data	Used to train Origin?
Your briefs, voice notes and brand rules	No, unless your business opts in in writing
Your product photos, footage and logos	No, unless your business opts in in writing
The ads Ori makes for you	No, unless your business opts in in writing
Faces and voices of people who gave likeness consent	Never. Used only for the scope they agreed
Feedback you choose to send us (for example, rating a cut)	Only if you opt in; otherwise used to fix the product, not to train
De-identified, aggregated usage measures (render times, error rates)	Used to improve the service, not to train on content

To opt in to helping improve Origin, or to withdraw an opt-in, write to info@zerohuman.co. Withdrawal applies to future training runs.

3. Brand memory#

Ori learns what works for your brand: from your onboarding answers, the rules you set, and the cuts you approve. Every belief is shown in Studio → Brand with how confident Ori is and where it came from. You can make Ori forget any belief, and remove any rule. Memory is per workspace, and is never used for, or visible to, any other customer.

4. Who can see your content#

You and your teammates in the workspace.
Our safety reviewers, only for briefs or ads that were flagged, to decide whether they break the Usage Policy.
Our support team, only when you ask for help with a specific ad or account.
Engineers, only through audited, just-in-time access to fix an incident, and never to browse.

5. Deletion#

Plans and rules you delete in the Studio are removed from the live system immediately, and beliefs you tell Ori to forget are no longer used. When you close a workspace, everything in it is deleted within 30 days, and from encrypted backups within a further 35 days. During the pilot, ask us to close a workspace or delete specific items at info@zerohuman.co; self-serve controls arrive before general availability. Full retention periods are in the Privacy Policy.

Privacy & dataPrivacy PolicyWhat personal data Zero Human AI collects when you use Ori and our website, why, who we share it with, how long we keep it, and the rights you have in the GCC, the EU and UK, and the US.AI & safetyResponsible AI & model governanceHow Zero Human AI builds, tests, releases and monitors Origin, the video model behind Ori: principles, release gates, evaluations, red-teaming, training-data governance and regulatory alignment.AI & safetyLikeness & Voice Consent PolicyWhen and how a real person’s face, body or voice can appear in ads made with Ori: identity checks, recorded consent, scope, revocation, and what’s never allowed.

Questions about this document? Write to info@zerohuman.co.