Course Content
Phase 1: The SIFT Protocol (Source Verification)
0/4
Lesson 1.1 The Neurology of the Pause
Lesson 1.2 The Four Moves: SIFT Deep Dive
Lesson 1.3: Domain Forensics & The Wayback Machine
Lesson 1.4: Tracking “Pink Slime” & Syndication Networks
Phase 2: Data Forensics & “Calling Bullshit”
0/10
Lesson 2.1 The Bad Science Audit
Lesson 2.2: The Fermi Estimation Check (Mental Math)
Lesson 2.3 Commercial Bias: The “Bad Pharma” Audit
Lesson 2.5 The Denominator Blindspot (How to Lie with Truth)
Lesson 2.6 Auditing Authority: The “Official Source” Trap
Lesson 2.7 The Autocrat’s Playbook: 20 Signs of Sliding
Lesson 2.8 The “Boring Truth” Doctrine
Lesson 2.9 The “Strategy of Tension”
2.10 The “Drug War” Smoke Screen: Narrative vs. Reality
2.11 The “Free Speech” Shield: Opinion vs. Incitement
Phase 3: Logic & Argument Analysis
0/3
Lesson 3.1: Propositional Logic (Beyond the Syllogism)
Lesson 3.2: The Principle of Charity (Steel-Manning)
Lesson 3.3: Cognitive Biases 2.0 (Motivated Reasoning)
Phase 4: Digital Investigation (The OSINT Toolkit)
0/7
Lesson 4.1: Reverse Image Search 3.0 (Yandex, Baidu, TinEye)
Lesson 4.2: Geolocation Basics (Mapping the Territory)
Lesson 4.3 Deepfake Detection (The Glitch List)
Lesson 4.4 (Part 2) – Video Forensics
Lesson 4.5 Lazy Journalism: The “Quote as Fact” Trap
Lesson 4.6 The Anatomy of a Disinfo Tweet
Lesson 4.7 THINK TANKS & THE “EXPERTISE” INDUSTRY
Phase 5:The Capstone (The Million Dollar Audit)
We tell the story of Sholto David not as a "news story," but as a Case Study in Tradecraft. He used the exact skills we just taught (Visual Forensics, Source Verification, Institutional Audit) to expose a massive lie and get paid for saving the taxpayer money.
0/1
5.1 The $2.6 Million Audit (Case Study: Sholto David)
TRADECRAFT: The Intelligence Analyst’s Guide to the Internet

 

TRADECRAFT: Video Forensics (The Dream Logic)

The Concept

Detecting fake video is harder than fake photos because the motion hides the errors. Your eye focuses on the main character (who usually looks perfect), while the background falls apart. AI video operates on Dream Logic: It forgets what an object is the moment it stops moving or passes behind something else.

THE CHECKLIST: The “Motion Tells”

1. Object Permanence Fails (The Vanishing Extra)

  • The Error: In real life, if a man in a red shirt walks behind a tree, he should emerge wearing a red shirt. In AI video, he might emerge wearing a blue suit, or turn into a mailbox.

  • The Check: Ignore the main character. Watch the background extras. Do they change clothes mid-walk? Do they vanish into thin air?

2. The “T-1000” Morph (Limbs vs. Objects)

  • The Error: AI struggles to separate foreground from background.

  • The Check: Watch when a person walks past a complex object (like a fence or another person).

    • Does their arm “melt” into the fence?

    • Do their legs briefly merge with the other person’s legs?

    • The Tell: Solid objects behaving like liquid.

3. The Shadow Flicker

  • The Error: AI often forgets to render shadows for every frame, or the shadows point in different directions as the camera moves.

  • The Check: Pick one shadow and stare at it. Does it pop in and out of existence? Does it detach from the person’s feet?

4. The “Lazy” Watermark

  • The Error: Many propagators of fake videos are lazy. They use free trials of generators like Kling, Luma, or Runway.

  • The Check: Look at the extreme corners of the video (especially Bottom Right).

    • Is there a blurred patch? (Where they tried to erase it).

    • Is there a faint logo that says “KLING” or “Runway”?

    • Pro Tip: Even if they crop it, the framing often looks “cramped” because they cut off the bottom edge to hide the logo.

LAB EXERCISE: The “Background Watch”

Task: Find a suspected AI video on X/Twitter.

  1. Play it once: Watch normally. It probably looks real.

  2. Play it again: Do not look at the main character. Pick a tree, a car, or a person in the distance.

  3. The Test: Track that background object for 10 seconds. If it morphs, floats, or disappears, the video is synthetic.

Previous
Next
Scroll to Top