I Tested 12 Ways to Spot Fake AI Video (They Work)
Understanding visual artifacts and physical flaws allows AI filmmakers to rigorously quality-control their own AI generations and maintain believable photorealism.
What this lesson covers
AI Samson demonstrates practical techniques to identify synthetic AI-generated video footage by analyzing recurring technical artifacts. His analysis highlights key breakdown areas such as object state changes, cyclical motion, eyelines, anatomy glitches, and broken physics.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 0:34 ↗
AI video models struggle with post-interaction object state changes, such as bitten food remaining whole or liquid levels failing to track pouring rate.
- 1:38 ↗
Cyclical motion artifacts, including background characters walking in perfect unison or environment elements looping rhythmically, expose synthetic video.
- 3:57 ↗
Unnatural eyelines, lack of mutual gaze between subjects, and erratic blink rates reveal desynchronized facial performance generation.
- 5:12 ↗
Anatomy glitches persist in small-resolution subjects, manifesting as morphing fingers, missing digits, or shifting dental geometry.
- 7:16 ↗
Complex physics violations, such as objects disappearing upon impact or unnatural trajectories, serve as immediate indicators of AI generation.
Workflow outlined in the video
- Perform state-change audits on generated footage to ensure objects and props respond accurately following character interactions.
- Inspect background crowds and environmental effects to eliminate repetitive looping or unnatural movement patterns.
- Verify character gaze vectors and blink frequencies across shots to maintain subject eyeline consistency.
- Zoom in on small-resolution details, specifically hands and teeth, to catch subtle geometry morphing or missing digits.
- Analyze fast-action physics in high-motion shots to confirm realistic collision mechanics and object trajectories.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
Explore the tools
How to interpret the numbers
22,812 views captured 24 October 2026. ReelStack recommendation score: 43/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 45% performance: 0.69× views vs 7 other channel videos observed at a similar age. View-evidence factor 96% (views / (views + 1,000))
- 25% freshness: 0/100 with a 14-day half-life
- 20% momentum: 74 views/day vs 18 channel baseline (5 comparable uploads), with the same view-evidence factor
- 10% engagement: 100/100; likes + 4× comments, smoothed with a 500-view neutral prior