[arXiv]score: 0.23

Vishing-Tactics-Bench Predicts Real-Time Harm Trajectories in Voice Phishing Calls

September 9, 2026

Vishing-Tactics-Bench evaluates the ability to forecast whether ongoing voice phishing calls will result in information gathering or financial exploitation. The benchmark includes 35,340 labeled scammer utterances from 5,645 synthetic calls to support real-time harm projection rather than post-hoc classification.

HOW THIS AFFECTS YOU

●

researcherYou can use this dataset to train models on trajectory forecasting rather than simple binary classification in high-stakes social engineering scenarios.

●

policyThis benchmark provides a framework for developing real-time defensive measures against voice-based fraud.

read original ↗arxiv.org

DAILY DIGEST

you don't check 9 sources — we do. one email every morning, read in 2 min. free. unsubscribe anytime. privacy

← back to feed