4 Ways to Fight AI-Based Fraud

This commentary outlines how cybercriminals are leveraging generative AI for deepfakes, biometric fraud, and prompt-injection to conduct convincing, high-value scams, citing cases like a $25M multiperson deepfake heist. It stresses that attackers pair AI with traditional social engineering (e.g., caller ID spoofing) and advises defenders to deploy AI/ML analytics for anomaly detection, strengthen anti-spoofing measures, improve data quality and telemetry (e.g., DPI) to ground AI models, embrace zero-trust, and baseline normal network behavior to rapidly spot anomalies.