"On 2 August 2026, Article 50 of the EU AI Act became applicable: AI-generated content must carry machine-readable marking, and substantially altered content a visible label. Most video teams cannot yet answer the auditor's first question: what exactly did AI do to this file?

This session is a live, end-to-end demonstration of a working answer. A single chat message triggers an AI clipping pipeline in which every operation is recorded and classified as assistive or substantive under the European Commission's guidelines. We show the side-by-side that matters in practice: one clip that legally requires no label, and one that does - with the reasoning for each. The system refuses to publish without a named human approval, applies the visible label only where required, and signs the output with C2PA Content Credentials chained back to the source master. We close by verifying the file live on stage with the open-source c2patool - so the audience does not have to take our word for anything - followed by Q&A.
"