Primary Investigator:
Quamar Niyaz

Project Members
Deeksha Hareesha Kulal

Phishing remains a critical cyber threat, with traditional ML-based detection models relying on grammatical errors and word anomalies as key indicators. However, LLM-generated phishing emails are well-structured, grammatically sound, and highly deceptive, making detection increasingly challenging. This research explores the impact of word correction and splitting techniques in strengthening ML-based phishing detection. We further investigate how these enhancements improve detection accuracy against well-crafted LLM-generated phishing emails and adversarial attacks, paving the way for more resilient and adaptive cybersecurity solutions.