AI-driven phishing increases click rates by threefold

Research indicates that the integration of artificial intelligence into cybercrime is significantly increasing the effectiveness of phishing attacks. A study from TU Berlin demonstrates that small, local large language models (LLMs) can personalize phishing emails so effectively that click rates increase by a factor of three.

Academic analysis further explores the role of machine learning and natural language processing in defending against these sophisticated threats. While AI systems excel at analyzing linguistic features to detect fraud, human decision-making remains highly susceptible to the psychological manipulation and emotional triggers used in modern phishing attempts. Current research suggests that a hybrid approach, combining AI-based detection with human oversight, may be the most effective strategy for mitigating these evolving risks.