2026 Undergraduate Research Showcase

BehavioVeri: A Multimodal Behavioral Authenticity Platform for Detecting External Assistance in Remote Hiring Assessments

Document Type

Student Presentation

Presentation Date

4-24-2026

Faculty Sponsor

Dr. Sindhu Kalathur

Abstract

Remote hiring interviews and online assessments face increasing integrity challenges as candidates leverage AI-generated content, external coaching, and digital assistance tools. Traditional proctoring systems rely on invasive environmental surveillance that raises privacy concerns while failing to detect sophisticated assistance strategies operating outside monitored devices. Current solutions cannot effectively identify AI-assisted responses delivered verbally or detect subtle forms of external guidance.

We have developed BehavioVeri, an anti-fraud platform that detects external assistance through behavioral authenticity analysis rather than environmental monitoring. The system employs a novel multimodal approach combining keystroke dynamics, facial micro-expressions, eye-gaze patterns, speech prosody analysis, and linguistic feature extraction to identify deviations from genuine human behavioral patterns. The platform's context-aware AI model performs real-time multimodal fusion to generate authenticity assessments. Unlike conventional proctoring tools, BehavioVeri operates using standard webcams, microphones, and keyboards without requiring specialized hardware or invasive data capture. What would the title be?

This document is currently not available here.

Share

COinS