Deception Detection¶

Why It Matters¶

In adversarial environments, sources have incentives to deceive. Accepting deceptive information as genuine leads to predictable errors. The challenge: deception is designed to be plausible, so it passes surface validity checks. Detection requires looking for structural signatures of deception rather than content assessment.

The Key Move¶

Ask: Who benefits from me believing this? Is the source in position to know this, or is it suspiciously convenient? Is the information too perfect - exactly what I'd want to hear, or what would confirm my hypothesis? Check for internal inconsistencies that wouldn't exist in genuine information. Look for "signals too strong" - information presented with implausible certainty or detail. Seek independent corroboration from sources without shared incentives to deceive.

Classic Application¶

Iraqi defector "Curveball" claimed Iraq had mobile bioweapons labs. His information was detailed and specific (signal too strong), confirmed U.S. suspicions (too convenient), couldn't be independently verified (no corroboration), and he had clear incentives (asylum, importance). These are classic deception signatures, but analysts failed to weight them against supportive content.

Surprising Application¶

Financial due diligence. When evaluating potential investments or acquisitions, financial statements that perfectly confirm investment thesis (margins in target range, growth at expected rates) warrant skepticism. Genuine businesses have irregularities, inconsistencies, and noise. Too-clean data suggests either exceptional luck or manipulation. One private equity firm flags "statistically improbable smoothness" in reported financials as a deception indicator.

Failure Modes¶

Assuming all sources are honest by default (adversarial environments require skepticism). Treating content quality as reliability signal (skilled deceivers provide high-quality false information). Dismissing everything as potential deception (paralysis - you must assess probability, not possibility). Not considering "deception to what end?" (understanding the deceiver's objectives reveals what they'd plausibly fabricate). Forgetting that people self-deceive (not all false information is deliberate lying).

Go Deeper¶

Whaley, "Stratagem: Deception and Surprise in War" on deception theory. Bennett and Waltz, "Counterintelligence: A Documentary History" Chapter 4. Heuer and Pherson, "Structured Analytic Techniques" Chapter 11.

Deception - Military Strategy & Strategic Thinking: A Map of the Territory