Case Law On Admissibility Of Ai-Powered Lie Detector Evidence
1. United States v. Scheffer (1998, U.S. Supreme Court)
Facts:
Sergeant Scheffer, in a military court‑martial, attempted to admit polygraph results as part of his defense.
Military Rule of Evidence 707 prohibited polygraph evidence in court‑martial proceedings.
Issue:
Can a categorical ban on polygraph evidence violate a defendant’s right to present a defense?
Holding:
The Supreme Court upheld the exclusion, ruling that the military rule did not violate the Sixth Amendment.
Reasoning:
Polygraph evidence is considered inherently unreliable: error rates vary, operators differ, and tests can be manipulated.
Military courts are allowed to make categorical evidentiary exclusions when reliability concerns are high.
Significance:
Establishes a precedent that courts are highly skeptical of lie-detector evidence.
Suggests that AI lie detectors would face the same scrutiny unless their reliability is rigorously demonstrated.
2. United States v. Cordoba (1998, C.D. California)
Facts:
The defendant sought to introduce polygraph results to support his defense.
The court held a Daubert hearing to determine admissibility.
Issue:
Does polygraph evidence satisfy the Daubert criteria for scientific evidence?
Holding:
Polygraph evidence was excluded.
Reasoning:
Fails Daubert factors: testability, peer review, known error rate, and general acceptance.
Risk of misleading the jury outweighed any probative value.
Significance:
Demonstrates the standard for scientific reliability that AI-powered lie detectors must meet for admissibility.
3. Selvi v. State of Karnataka (2010, Supreme Court of India)
Facts:
Considered the use of narco-analysis, polygraph, and brain-mapping tests in criminal investigations.
Petitioners challenged involuntary administration as a violation of rights.
Issue:
Can such tests be compelled, and can their results be admitted in court?
Holding:
Involuntary tests violate Articles 20(3) (protection against self-incrimination) and 21 (right to life and liberty).
Test results cannot be used as evidence unless conducted voluntarily and under strict safeguards.
Reasoning:
Protects personal liberty, bodily integrity, and mental privacy.
Significance:
Sets strong procedural and constitutional safeguards, highly relevant to AI-based lie detectors.
Consent and voluntariness are crucial.
4. Wolfel v. Holbrook (6th Circuit, 1987)
Facts:
In a civil case, the plaintiff submitted polygraph results related to alleged misconduct by a corrections officer.
Issue:
Are polygraph results admissible without prior agreement on methodology or foundation?
Holding:
Polygraph evidence was inadmissible.
Reasoning:
Requires advance stipulation or proper procedural foundation.
Without such foundation, the evidence could unfairly influence credibility assessments.
Significance:
Reinforces the requirement for strict procedural and methodological standards—critical for AI lie-detector evidence.
5. Ashwini Kumar Upadhyay v. Union of India (Delhi High Court, 2023)
Facts:
Petition challenged proposals to force complainants or accused to undergo polygraph, narco-analysis, or brain-mapping tests.
Issue:
Can such tests be made compulsory for establishing credibility?
Holding:
Tests cannot be compelled; voluntary consent is required.
Reasoning:
Forced tests violate dignity, personal liberty, and self-incrimination protections.
Significance:
Highlights that even voluntary AI-powered deception detection must respect individual rights.
Key Takeaways for AI-Powered Lie Detector Evidence
Reliability and Validation – Courts demand rigorous proof of accuracy and standardisation.
Consent – Voluntary participation is mandatory; forced testing is generally inadmissible.
Procedural Safeguards – Proper foundation, operator qualifications, and error-rate disclosure are essential.
Skepticism – Courts treat lie-detector evidence with caution; AI evidence will be similarly scrutinized.
Comparative Jurisdictions – U.S. emphasizes scientific reliability (Daubert), India emphasizes constitutional rights.
RELATED Blog
comments