AI-based phishing detection focuses on identifying deceptive digital content designed to mislead users into interacting with malicious links, QR codes, or email messages. It involves extracting relevant indicators from QR-code destinations and email content, analyzing suspicious URLs and linguistic patterns, and evaluating features that distinguish legitimate communications from potential phishing attempts. By combining image-based classification with transformer-based natural language analysis, this approach helps identify phishing-related characteristics across different communication formats and classify potentially malicious content for further investigation. The resulting detections support early identification of phishing threats, improved security awareness, and the implementation of appropriate protective measures to reduce the risk of credential theft and unauthorized access.