Episode 6 AI-video prompt — A4 report discovery

Use this as a photographic motion plate. Composite the exact report text and highlighted words afterward; generative video should not be trusted to spell them correctly.

Prompt

Create a restrained cinematic 16:9 shot, 8 to 10 seconds, 4K, 24 fps source motion for a serious near-future technology thriller. Night interior of a regulated bank security operations room. A single realistic white A4 security report lies on a dark charcoal desk beneath a cool overhead task light. A young Asian woman investigator in a black blazer is present only as soft out-of-focus hands and silhouette at the edge of frame, maintaining continuity with Sam; do not reveal a new face. The camera begins in a slightly high three-quarter view and performs an extremely slow, stable push toward the paper with subtle natural parallax. Cyan monitor reflections move faintly across the desk. Four separate locations in the report acquire a restrained warm amber scanning glow, one after another, as though document-analysis software has detected four unusual words. Fine amber guide lines extend from those four locations toward clean negative space on the right side of frame. The feeling is discovery, not magic: credible forensic analysis, quiet tension, premium streaming-drama cinematography, realistic paper fibres, shallow but sufficient depth of field, soft film grain, controlled highlights, deep blacks, no handheld shake.

Leave the rightmost 35 percent of the frame dark and visually uncluttered for a later typography overlay. Keep the full A4 page inside frame and mostly front-facing so tracked graphics can be composited accurately. Do not generate readable report prose, interface text, floating letters, holographic screens, logos, watermarks, subtitles, dramatic lens flares, fast zooms, glitch effects, or camera rotation. Do not make the paper itself glow. End on a steady two-second hold after all four amber locations have been marked.

Exact composite added afterward

The four highlighted words, in order, are:

AMOUNT · REFERENCE · TIME · NEXT…

On the right, assemble them under:

EXTRACTED WORDS · READING ORDER

Then add:

INCOMPLETE MESSAGE · ONE REPORT IS NOT ENOUGH

Suggested edit

• 0.0–1.5 s: establish the paper and Sam's presence.
• 1.5–5.5 s: mark the four word positions one at a time.
• 2.0–6.0 s: composite exact word copies travelling along the guide lines.
• 6.0–8.0/10.0 s: hold the completed four-word fragment for reading.

If the video model cannot preserve a stable sheet of paper, ask it for the same shot with a blank A4 page and no glowing marks. The deterministic HTML layer can provide the entire report, highlights, extraction lines, and exact message.