VEXA: Evidence-Grounded and Persona-Adaptive Explanations for Scam Risk Sensemaking

2026-02-04

AI Summary

This work addresses the limited transparency and misalignment with decision logic in existing fraud detection explanations, which hinder non-expert users’ everyday risk assessment. To bridge this gap, the paper proposes VEXA, a novel framework that integrates GradientSHAP-based evidence attribution with psychologically informed user vulnerability profiles to generate semantically faithful and stylistically adaptable explanations. Guided by the principle that “semantics are determined by evidence while expression is modulated by user role,” VEXA is evaluated on multi-channel fraud data and demonstrates significantly improved semantic reliability without increasing linguistic complexity. Moreover, it enables role-specific explanatory styles tailored to diverse user personas while preserving consistency in underlying evidence attribution.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Knowledge Representation and Reasoning: Preferences
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Online scams across email, short message services, and social media increasingly challenge everyday risk assessment, particularly as generative AI enables more fluent and context-aware deception. Although transformer-based detectors achieve strong predictive performance, their explanations are often opaque to non-experts or misaligned with model decisions. We propose VEXA, an evidence-grounded and persona-adaptive framework for generating learner-facing scam explanations by integrating GradientSHAP-based attribution with theory-informed vulnerability personas. Evaluation across multi-channel datasets shows that grounding explanations in detector-derived evidence improves semantic reliability without increasing linguistic complexity, while persona conditioning introduces interpretable stylistic variation without disrupting evidential alignment. These results reveal a key design insight: evidential grounding governs semantic correctness, whereas persona-based adaptation operates at the level of presentation under constraints of faithfulness. Together, VEXA demonstrates the feasibility of persona-adaptive, evidence-grounded explanations and provides design guidance for trustworthy, learner-facing security explanations in non-formal contexts.

Problem

• online scams
• explainable AI
• risk sensemaking
• persona-adaptive explanations
• evidence-grounded explanations

Innovation

• evidence-grounded explanation
• persona-adaptive
• GradientSHAP
• scam detection
• interpretable AI