We disclose what we make.
Synthetic media only stays useful while audiences can still trust what they are shown. These are the rules we work to — on client campaigns and on our own products alike.
- 01
AI content disclosure
Work we produce using generative tools is identified as such. Disclosure is a default, not a concession made when someone asks.
- Any delivered asset whose imagery, motion, voice or performance is generated is labelled AI-generated in the delivery record, and we supply disclosure copy suitable for the placement.
- Where a platform provides a synthetic-media flag, we ask the client to enable it and we tell them where it is.
- Where a jurisdiction requires machine-readable marking of AI-generated content, we support embedded provenance signals rather than relying on a caption alone.
- We do not remove, obscure or advise removing a disclosure to improve performance.
- 02
Synthetic talent & consent
A generated person is not a free person. Likeness and voice carry rights whether or not the individual appears on set.
- We do not generate a recognisable likeness of a real, identifiable individual without that person's documented written consent for the specific use.
- We do not clone a voice without documented written consent from the voice's owner for the specific use.
- We do not generate depictions of a client's actual named employees, customers or executives unless the client supplies that written consent.
- Fully synthetic presenters who do not depict a real individual are permitted, and are disclosed as synthetic under policy 01.
- Consent records are retained with the campaign's provenance record.
- 03
Claim substantiation
Creative can make a product look good. It cannot make a claim true. Substantiation is the client's evidence and our checklist.
- Any performance, health, financial, safety or comparative claim in a script must be backed by evidence the client supplies in writing before production.
- We do not script or produce testimonials, reviews, or first-person experience claims from people who did not have that experience. No exceptions, no framing that makes it acceptable.
- A synthetic presenter does not deliver first-person claims of personal use or personal results, because no person made them.
- Health and financial subject matter receives a stricter pass: unsubstantiated claims are removed rather than softened.
- If a client will not substantiate a claim, we produce the work without it or decline the brief.
- 04
Rights, licensing & provenance
Every delivery carries a record of what it was built from, so the client can answer questions about their own campaign months later.
- Source imagery must be owned by the client or properly licensed. We confirm the basis before production, not after a complaint.
- Music and stock assets are licensed for the actual distribution, including paid media where applicable.
- Model and tool licence terms are checked for commercial-use permission for the specific deliverable.
- Each delivery includes a provenance record: source assets, generative tools used, human decisions applied, consents obtained, and the disclosure text supplied.
- Records are retained so a campaign can be reconstructed and defended after the fact.
- 05
Campaign disclosure guide
Practical guidance we hand to clients, because a policy that only exists on our side does not protect anybody.
- Sponsored and partnership content carries the platform's own disclosure in addition to any AI labelling — the two requirements are separate and both apply.
- Disclosure belongs where it is actually seen, not buried at the end of a caption or in a comment.
- Where a campaign runs across jurisdictions, we apply the strictest applicable standard to the whole campaign rather than maintaining divergent versions.
- If a client insists on a placement that removes required disclosure, we do not deliver that version.
Two regimes, both live.
United States — FTC
- Where AI created or substantially modified advertising content, that use should be disclosed to consumers. Grammar checking and analytics tooling are not what this is aimed at; AI-generated imagery, video, voice and ad copy are.
- Sponsored content carries two separate obligations: the paid relationship AND the AI involvement. Satisfying one does not satisfy the other.
- Disclosure must be clear and conspicuous — placed where it is actually seen. For video, that means on-screen and early rather than buried in a caption.
- AI does not launder a claim. Anything asserted in generated content must be truthful and substantiated exactly as if a person had said it.
- Penalties are assessed per violation, and each non-compliant post can count separately — which is what makes a large campaign the expensive way to learn this.
European Union — AI Act, Article 50
If you run campaigns into the EU, this is not a future problem to schedule. It is days away at the time of writing.
- Deployers creating deepfake content must disclose that it has been artificially generated or manipulated. The definition is broad: content resembling real persons, objects, places or events that would appear authentic.
- Content that looks or sounds like a real person requires labelling even where no deception was intended — and, on the current guidance, even where no specific real individual is depicted.
- Providers of systems generating synthetic audio, image, video or text must ensure outputs are marked in a machine-readable format and detectable as artificially generated. A visible caption alone does not satisfy this; provenance signalling matters.
- Penalties reach into the millions of euro or a percentage of worldwide annual turnover, whichever is higher.
How we operate against this. Where a campaign touches more than one jurisdiction we apply the strictest applicable standard across the whole campaign rather than maintaining divergent versions — it is cheaper than segmenting, and it removes an entire category of mistake. Exact figures, thresholds and effective dates change; confirm your own position with qualified counsel rather than relying on this page.
We do not fabricate people or proof.
No invented testimonials. No fabricated reviews. No synthetic person claiming an experience they did not have. No unsubstantiated health or financial claim. These are not preferences to be traded against a deadline or a budget — a brief that requires any of them is a brief we turn down.
Scope. This is Quantivo AI's own operating policy, published so clients and audiences can hold us to it. It describes how we work; it is not legal advice, and it does not replace a client's own regulatory obligations in their market. Advertising, disclosure and AI-labelling requirements differ by jurisdiction and change over time — clients should confirm their specific obligations with qualified counsel. Where our policy and a client's legal requirement differ, the stricter standard applies.