Public Research Artifacts

The following materials are released publicly to establish shared language around perceptual alignment in voice systems. They are not comprehensive and are not intended to replace direct engagement.

 

 ________________

 

 

Tonality as Attention

White Paper | Zenodo  | October 2025 (590+ downloads)

 

A conceptual framework examining how human vocal tonality functions as an attention-shaping and trust-modulating signal in voice systems.

This paper formalizes a class of perceptual failures many teams experience intuitively but struggle to articulate or measure.

 

Intended for:
Research leaders, product teams, and executives seeking clarity around felt experience, trust, and voice-mediated interaction.

View on Zenodo - Tonality as Attention: Bridging Human Voice Tonality and AI Attention Mechanisms to Reintroduce the Human Layer to Intelligence (DOI 10.5281/zenodo.17410581)

 

 ________________

 

 

TonalityPrint

Single-Speaker Tonal Dataset | Zenodo | January 2026

 

A high-control, single-speaker dataset designed for perceptual calibration, interpretability, and continuity across rapidly evolving voice models.

TonalityPrint is not optimized for scale or diversity.
It exists to support controlled investigation of tonal causality and trust stability. 

 

Strategic Implementation Intended for:
Organizations shipping  voice to humans who require a safe, stable & responsible human reference point across model changes.

 

The dataset architecture and technical documentation are archived on Zenodo for peer review and provenance

 

View on Zenodo TonalityPrint: A Contrast-Structured Voice Dataset for Exploring Functional Tonal Intent, Ambivalence, and Inference-Time Prosodic Alignment  .DOI 10.5281/zenodo.17913895

 

 

 ________________

 

 

 

Executive Overview

Perceptual Alignment for Voice AI

 

A concise, executive-facing overview outlining why perceptual instability emerges in modern voice systems - and why fine-tuning alone does not reliably resolve it.

This overview is designed to support rapid internal alignment and decision-making.

 

Intended for:
CEOs, VPs, and senior research or product leaders navigating voice AI's tonality risk under time pressure.

 

 

Prepared to support rapid internal alignment under perceptual uncertainty

 

Download  Executive Overview Perceptual Alignment Voice AI 

 

 

 ________________

 

Private Engagement

Public artifacts represent a small subset of the work.

 

 

Organizations engaging privately typically do so when:

     -Voice AI systems sound fluent but feel subtly wrong.

     -Teams sense tonal trust drift across model upgrades.

     -Demos succeed technically but fail perceptually.

     -Leadership needs clarity before shipping or scaling.

     -Private work focuses on perceptual tonal intent alignment rather than model performance alone.

 

                           ________________

 

 

Engagement Structure

Engagements begin with a Confidential Strategic Perceptual Briefing. Further access is offered selectively.

 

The private briefing establishes:

   -Relevance and urgency

   -Scope of perceptual risk

   -Appropriate next steps, if any

 

  Subsequent access may include:

   -Perceptual audit sprints

   -Tonal reference dataset licensing

   -Embodied human voice reference licensing

   -Ongoing strategic advisory

 

Availability is intentionally limited.

 

  

 

 

Strategic Access

If you are building voice systems for humans and are encountering - or suspecting - perceptual tonal intent instability, our conversation is time-sensitive.

 

 Request Confidential Strategic Access

Requests are reviewed personally.