VoiceGuard: An Evaluation-Honest, Multi-Signal Decision Layer for Real-Time Voice-Clone Detection in High-Stakes Telephone Calls
- 1Vidyalankar Institute of Technology, Mumbai, India
Abstract
VoiceGuard: An Evaluation-Honest, Multi-Signal Decision Layer for Real-Time Voice-Clone Detection in High-Stakes Telephone Calls
This paper investigates the generalization limitations of speech anti-spoofing systems for detecting voice-clone fraud in telephone calls. It examines evaluation artifacts in the ASVspoof 2019 Logical Access benchmark and re-evaluates the published AASIST model using an anti-shortcut evaluation protocol.
The paper proposes VoiceGuard, a multi-signal decision layer that combines acoustic spoof detection, speaker consistency, prosodic features, and call context to generate ALLOW, VERIFY, or ESCALATE recommendations. It also introduces a feature-only audit log to support transparency and human review.
Additionally, a five-language pilot study evaluates the transferability of an English-trained detector across Indian languages, highlighting the challenges of multilingual voice-clone detection.
The work emphasizes evaluation integrity, cross-dataset generalization, explainable decision-making, and the need for further validation before deployment in high-stakes environments.
Authors: Ghanshyam Kumavat, Sunil Saini, and Vicky Pukale
Affiliation: Department of Computer Engineering, Vidyalankar Institute of Technology, Mumbai, India.
Related
- Sharp Limits for Honest Uncertainty in Hard-Budget Repeated Evaluation
- Evaluation of clinical interventions made by pharmacists, doctors and nurses in a hospital post-discharge telephone service: a retrospective cohort study
- Foundation models for tabular medical detection: a robust and generalizable framework for clinical decision support
- Reliable Detection, Unreliable Clinical Decision-making: Artificial Intelligence Models in Burn Assessment
- F2D-RetinaNet: a feature-to-decision dual-path network with fractional operators for retinal disease detection and grading
BibTeX
@misc{kumavat2026voiceguard,
title = {{VoiceGuard: An Evaluation-Honest, Multi-Signal Decision Layer for Real-Time Voice-Clone Detection in High-Stakes Telephone Calls}},
author = {Ghanshyam Kumavat and Sunil Saini and Vicky Pukale},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23451734},
url = {https://zenodo.org/records/23451734},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - VoiceGuard: An Evaluation-Honest, Multi-Signal Decision Layer for Real-Time Voice-Clone Detection in High-Stakes Telephone Calls
AU - Ghanshyam Kumavat
AU - Sunil Saini
AU - Vicky Pukale
PY - 2026
DA - 2026/10/18
PB - Zenodo
DO - 10.5281/zenodo.23451734
UR - https://zenodo.org/records/23451734
AB - VoiceGuard: An Evaluation-Honest, Multi-Signal Decision Layer for Real-Time Voice-Clone Detection in High-Stakes Telephone Calls This paper investigates the generalization limitations of speech anti-spoofing systems for detecting voice-clone fraud in telephone calls. It examines evaluation artifacts in the ASVspoof 2019 Logical Access benchmark and re-evaluates the published AASIST model using an anti-shortcut evaluation protocol. The paper proposes VoiceGuard, a multi-signal decision layer that combines acoustic spoof detection, speaker consistency, prosodic features, and call context to generate ALLOW, VERIFY, or ESCALATE recommendations. It also introduces a feature-only audit log to support transparency and human review. Additionally, a five-language pilot study evaluates the transferability of an English-trained detector across Indian languages, highlighting the challenges of multilingual voice-clone detection. The work emphasizes evaluation integrity, cross-dataset generalization, explainable decision-making, and the need for further validation before deployment in high-stakes environments. Authors: Ghanshyam Kumavat, Sunil Saini, and Vicky Pukale Affiliation: Department of Computer Engineering, Vidyalankar Institute of Technology, Mumbai, India.
ER -