AI-generated phishing attack detection focuses on identifying deceptive messages and phishing content created or enhanced using artificial intelligence to make fraudulent communications appear convincing and legitimate. It involves analyzing message semantics, linguistic patterns, contextual meaning, and suspicious communication characteristics to distinguish phishing attempts from legitimate messages across web-based collaboration platforms. By combining large language model-based semantic feature extraction with machine learning classification techniques, this approach helps identify AI-generated phishing content, recognize multilingual phishing messages, and detect deceptive language patterns that may bypass conventional keyword-based filters. The resulting findings support early phishing identification, improved detection of language variations and sophisticated social engineering techniques, reduced risk of credential theft and unauthorized access, and the implementation of appropriate email and collaboration-platform security controls.