RENEE ELLORY


Truth Stress Testing · Deception Detection · Adversarial Reasoning

I break open lies, evasions, and polished nonsense—fast.

I’m an independent consultant applying rare forensic deception-detection skill and adversarial thinking to expose false confidence, hidden motives, and system failures—in people, institutions, and AI systems.

This is not advisory work.
This is truth enforcement under pressure.

What I Do

I take narratives, decisions, or systems that appear sound—and force them to prove they’re real.

If they fail, I identify exactly where, why, and what collapses next.

Core Capabilities

  • Surgical identification of lies, evasions, and internal contradictions

  • Hallucinations, guardrails, and “smooth falsehood” detection in AI systems

  • Incentive-misalignment and failure-mode stress testing

  • High-stakes decision audits where error is irreversible

Where I’m Most Useful

You should contact me when:

  • The explanation is polished—but something feels off

  • Confidence is high and evidence is thin

  • An AI system is persuasive, but truth-faithfulness is necessary

  • A decision carries legal, ethical, or reputational risk

  • Everyone agrees—and that’s exactly the problem

Areas of Work

  • AI Truth-Stress Testing
    Adversarial testing of frontier models for hallucinations, evasions, refusal integrity, and incentive distortion.  I surface failure modes that pass internal benchmarks but fail reality.
  • High-Stakes Decision Audits
    For deals, disputes, institutional failures, and sensitive organizational systems—where one wrong assumption cascades.
  • Deception & Credibility Analysis
    Pattern-based detection of inconsistencies, linguistic leakage, lack of coherence, and behavioral misalignment.

Credentials (Signal, Not Theater)

  • Verified participant in the Ekman / O’Sullivan “Wizards Project”, identifying outlier deception-detection accuracy (~top 0.33%)

  • Decades of applied behavioral pattern recognition under real-world stakes

  • Sustained, top-percentile adversarial interaction with frontier AI systems driven by forensic curiosity

How I Work

  • Short, high-intensity engagements

  • NDA-first, highly selective

  • The goal is closure and accuracy—not comfort.

What You’ll Get

  • Claims separated cleanly from evidence

  • Hidden assumptions exposed

  • Failure paths identified

  • Clear constraints named

Fit Matters

I am not a fit if you want:

  • Validation

  • Comfort

  • Polished explanations

  • Someone to make a decision feel easier

I am a fit if you want:

  • The truth, even if it’s inconvenient

  • To know what breaks before it breaks publicly

Engagements

Work is highly selective and NDA-protected.

If you believe your qsystem, or decision must not be wrong, you may request a screened inquiry.