A Survey on Large Language Models in Phishing Detection

Phishing attacks have affected the society more than ever, therefore, detecting such attacks has been a significant challenge. This research examines the potential use of large language models (LLMs) in phishing attacks and their detection capabilities. In this context, various researches have been analyzed that explain how LLMs can be used to generate different phishing techniques, such as fake websites, emails, and text messages. In addition, the analyzed researches present various approaches to assess whether the language processing and text analysis capabilities of LLMs are effective in detecting phishing attacks. The analyses results show that LLMs are powerful tools for both attacking and developing defense mechanisms related to phishing attacks.