How AI Slop Poisons the Evidence Trail
A colorful animated explainer turns a flood of low-quality media into a concrete case study of how unsupported claims become apparent sources; metrics cover the first six minutes of a longer upload.
Field notes / why it works
The opening claim starts at 0:00 while mismatched image cards multiply through the first three seconds, giving the abstract problem an immediate visual shape. The busiest example collage lands at the highest replay peak, 0:29.6–0:35.5. A concrete test starts at 2:44, a plausible partial success raises the question at 3:25, and the 5:30–5:47 payoff shows how an unreliable claim can circulate as apparent evidence.
Plate 01 / The format in one breath
A calm, quick narrator makes a broad threat tangible through a colorful flood of examples, then tests the problem through a specific research story. The argument escalates from a surprising result to a feedback loop, with playful illustrated characters and interfaces keeping an abstract topic visible.
Why it works
- The first sentence states the concern at 0:00 while an empty black frame fills with increasingly odd image cards between 0.25 and 3.00 seconds. The picture poses the question before the explanation begins.
- The early accumulation of examples culminates in the most replayed segment, 0:29.6–0:35.5: a dense collage of absurd content, then a sample content grid around 0:36.5.
- At 2:44 the film switches to a concrete test; at 3:16 it gives a plausible partial success, then at 3:25 opens the more important question about the remaining claims. The payoff at 5:30 shows how a weak claim can reappear as apparent evidence.
Plate 02 / Format card
| Measured six-minute segment | |
|---|---|
| Platform · aspect · length | YouTube · 1280×720, 16:9, 60 fps · 6:00 analyzed |
| Pace | 35 detected shots, 34 definite cuts, 5.7–13.0 possible cuts/min with uncertain flashes/moves; 48 internal graphic steps, 13.7 definite picture changes/min; mean shot 10.29 s, median 7.08 s |
| Script | 986 transcribed words; 164 wpm overall, 190 wpm during speech; 52 sentences averaging 19 words; two questions; little direct address |
| Voice | One low-register, animated, measured adult narrator; explanatory and lightly dry |
| On screen | Approximately 65% illustrated character/world scenes, 20% interface or document graphics, 15% collages and abstract systems (visual estimate from sheets); no presenter footage |
| Captions | No continuous subtitles in sampled frames; selective all-caps headlines, diagrams and interface labels |
| Sound | Continuous tonal bed measured about 11 dB beneath voice; cut accents are suggested by loudness changes on 79% of detected cuts; −15.6 LUFS integrated, −3.1 dBFS true peak |
| Structure | Threat → example barrage → why it matters → research process → test → suspect claims → source tracing → real-world spread; no final CTA visible in measured segment |
Pro / 13 plates + recipe
The first plates are open. The full dissection is Pro.
Get the timed shots, type and sound specs, and the runnable recipe.
Open the full blueprint- Plate 03 / The hook, frame by frame (first 3 seconds)Pro
- Plate 04 / Structure (beat sheet)Pro
- Plate 05 / Shot grammarPro
- Plate 06 / Visual style guidePro
- Plate 07 / Voice and deliveryPro
- Plate 08 / Sound designPro
- Plate 09 / PackagingPro
- Plate 10 / Script templatePro
- Plate 11 / Production planPro
- Plate 12 / PromptsPro
- Plate 13 / Edit recipePro
- Plate 14 / QA checklistPro
- Plate 15 / Make it yoursPro
- Recipe / runnable promptPro