Usage Policy

1. Who this applies to#

This policy applies to everyone who uses Ori, the Origin API or anything made with them, including through agencies, integrations or teammates. It’s part of our Terms of Service. It also applies to the ads you publish: making an ad with Ori and then using it in a way this policy forbids is a breach.

Ori also follows the advertising policies of the platforms it publishes to, and the advertising laws of each market. When a platform or a law is stricter than this policy, the stricter rule applies.

2. What Ori won’t make#

Don’t use Ori to create, or try to create:

Elections and politics

• ads for or against political parties, candidates, elected officials, ballot measures, referendums or elections;
• ads on contested political or social issues, or that aim to influence voting or civic participation;
• content that impersonates governments, public bodies or officials.

Real people without consent

• any real, identifiable person, by face, body or voice, unless they’ve completed our consent flow for that use;
• lookalikes or soundalikes of real people, celebrities or public figures, or content that implies their endorsement;
• anyone under 18 as a real person; or anyone in a sexual, demeaning or compromising situation.

Deception and fraud

• false or unsupported claims about health, results, savings, earnings, safety or the environment;
• fake reviews, testimonials, endorsements, awards or certifications; fake before-and-after results;
• fake news reports, interviews or documentary footage, or realistic depictions of events that didn’t happen, presented as real;
• scams, fake giveaways, phishing, pyramid schemes, fake scarcity and other deceptive offers;
• ads for brands, products or businesses you don’t own or represent.

Harmful content

• child sexual abuse material, or any sexualisation of minors. We report it to the authorities;
• sexually explicit content or nudity, and non-consensual intimate imagery;
• hate, harassment or demeaning stereotypes based on protected characteristics such as race, ethnicity, religion, nationality, sex, gender, sexual orientation, age or disability;
• violence, terrorism or violent extremism, or content that glorifies or incites them;
• promotion of self-harm, eating disorders or dangerous challenges.

Illegal or restricted products

• illegal drugs, weapons, explosives, counterfeit goods, or anything illegal where the ad will run;
• alcohol, tobacco, vaping, pork or gambling in markets that ban those ads, or to audiences who can’t legally buy them;
• prescription medicines, unapproved supplements or medical services where advertising them is restricted.

Misusing the service

• removing or obscuring content credentials, watermarks or platform AI labels;
• evading our checks, for example by splitting a request, using coded language or creating new accounts after a suspension;
• using Ori or its outputs to train or build a competing model, or to scrape, attack or overload any service;
• infringing copyright, trademarks or other rights (see the Copyright & IP Policy).

3. Ads that need extra care#

These categories are allowed, but Ori treats them as high-risk. For them, Ori always asks before posting, whatever autonomy mode you choose, and:

We may ask for evidence behind a claim, and may pause an ad until we receive it.

4. Being open about AI#

• Keep the content credentials and invisible watermark Ori adds to every export.
• When Ori posts for you, it switches on the platform’s AI label. If you post yourself, switch it on too where the platform offers one.
• If an ad shows realistic people, places or events that didn’t exist or didn’t happen, make that clear to viewers, for example with Ori’s optional “Made with AI” mark. In the EU and some other markets this is a legal duty for deepfakes, and it’s the advertiser’s duty.
• Don’t present an AI-generated spokesperson as a real customer or employee.

5. How we enforce this policy#

Each brief and ad passes five checks: the brief, the script, the frames, the voice track, and, when something is flagged, a trained person. See Safety & Enforcement.

1. A blocked request isn’t, on its own, a violation. You’ll see why it was blocked and what you can change.
2. Repeated or deliberate attempts lead to a warning, then a temporary suspension, then termination.
3. Serious harms, such as child safety, non-consensual imagery, fraud or impersonating a real person, lead to immediate suspension and, where the law requires, a report to the authorities.

To appeal a decision, reply to the notice or write to info@zerohuman.co. A different reviewer answers within 5 business days.

6. Reporting misuse#

If you see an ad made with Ori that breaks this policy, or that uses your face, voice or work without permission, write to info@zerohuman.co with a link or a screenshot. For content that uses your likeness without consent, we take it down first and investigate after.

Version history