Lie Detector Alternative:
What Works, What Doesn't, and Why

No technology detects lies. What polygraph, EyeDetect, voice stress analysis, and FACS-based systems like EchoDepth actually measure are physiological correlates of stress, suppression, and cognitive load — signals that correlate with deception attempts but are not definitional proof of deception. Understanding this distinction is essential for evaluating any credibility assessment technology.

Why "Lie Detector" Is a Misnomer

The term "lie detector" implies that a technology can directly identify when someone is lying. No such technology exists. What credibility assessment tools measure are indirect physiological signals that research has found to correlate with deception in group-level studies. Individual-level prediction is less reliable, particularly for motivated, trained, or pathological subjects.

The distinction matters for procurement: a system that "detects deception" makes a claim that cannot be fully substantiated. A system that "surfaces involuntary physiological correlates of stress and suppression" makes a verifiable, defensible claim — which is the basis on which EchoDepth's credibility assessment is positioned.

Polygraph: The Original Lie Detector Alternative Problem

Polygraph — still the most widely deployed credibility assessment technology — measures respiratory rate, skin conductance, and cardiovascular response as proxies for anxiety during questioning. The US National Academy of Sciences concluded in 2003 that polygraph lacks scientific validity for security screening, with false negative rates reaching 47% in controlled studies. Despite this, many organisations still rely on it for lack of a better-evidenced alternative.

Polygraph's core problem is specificity: elevated physiological arousal is produced by many states other than deception — anxiety about being tested, embarrassment, surprise, genuine anger. Countermeasures (physical tension, mental arithmetic during control questions) can suppress the differential that examiners use to make assessments.

Voice Stress Analysis (VSA): No Scientific Basis

Voice Stress Analysis claims to detect deception through micro-tremors in speech. Multiple independent scientific assessments — including a 2002 US Department of Defense Polygraph Institute study — found that VSA performs at chance level in controlled conditions. It has no peer-reviewed validation and is not accepted in any legal or intelligence procurement context that applies evidential standards.

EyeDetect: One Channel, Controlled Conditions Required

EyeDetect (Converus) measures pupillary response during a structured computer-based test. It monitors one physiological channel — pupil dilation and eye movement patterns — and requires the subject to complete a specific test protocol. Peer-reviewed validation is limited, results require cloud processing on US servers (disqualifying it from SCIF deployment), and the test format is impractical for continuous or natural-interaction monitoring.

FACS-Based Credibility Assessment: The Evidence-Based Alternative

FACS-based credibility assessment — as implemented in EchoDepth — analyses 44 facial Action Units per frame, including micro-expressions, suppression patterns, and AU temporal sequences that are physiologically constrained and difficult to consciously control. Key differences from polygraph and VSA:

• 44 channels simultaneously vs polygraph's 3–4 peripheral physiological signals
• Camera-only, non-contact vs physical sensor attachment
• Structured, timestamped JSON output vs subjective examiner report
• SCIF and air-gap compatible vs cloud or examiner dependency
• Peer-reviewed 40-year foundation (FACS) vs contested or absent evidence base

EchoDepth does not claim to detect lies. It surfaces involuntary facial physiological correlates — stress, suppression, micro-expressions — that peer-reviewed FACS research has associated with deception attempts. The output is a structured, reproducible, legally reviewable record.

Choosing a Lie Detector Alternative for UK Defence

For UK defence, intelligence, and security procurement, the key evaluation criteria are: scientific validity (peer-reviewed evidence base), legal defensibility (structured auditable output), operational practicality (non-contact, works during natural interactions), sovereignty (UK data residency, SCIF-compatible), and integration (SIEM, DSAT compatibility).

EchoDepth meets all five criteria. Polygraph meets none definitively. EyeDetect meets scientific validity partially but fails on sovereignty and operational practicality. VSA meets none.