How we handle AI
AI lets us build things that would otherwise be impossible to film. Here is how we keep that honest.
Every line below describes something we do on every project. No one certifies it.
The transparency obligations these practices anticipate apply from 2 August.
We label AI-generated work.
We disclose AI-generated and AI-modified content in our deliverables, in line with the transparency obligations in Article 50 of the EU AI Act.
In every deliverable.
Every project comes with an AI-use declaration.
You get a short written record with each delivery: which elements were AI-generated, which were AI-modified, and which are camera-original.
With each delivery.
Provenance stays intact.
Our tools embed provenance data in what they produce. That data survives our edit and our export, and a visible label goes on the content itself.
Through the edit and the export.
Consent covers synthetic use.
Where we use a real person’s likeness or voice, we get written, revocable consent that covers synthetic reproduction. The consent exists before the likeness is built.
Before anything is built.
A person reviews the science.
Medical and patient-facing material passes named human review before publication. AI-generated health copy is released only after that review.
Before anything is published.
We stay current.
Our practice tracks the AI Act’s transparency rules and the guidance issued under them.
As the guidance changes.
If you want to know how any of this would work on your project, ask us.