This research evaluates the effectiveness of LLM-generated phishing emails against traditional human-crafted attacks in large organizational settings.

• LLMs can create highly personalized, error-free phishing content at scale
• The study provides comparative analysis of detection rates and user susceptibility
• Findings reveal specific vulnerabilities in current security protocols against AI-generated attacks
• Researchers propose defensive strategies to mitigate these emerging threats

For cybersecurity leaders, this research highlights the urgent need to update security training and technological defenses to address the growing sophistication of AI-enabled social engineering attacks.

Lateral Phishing With Large Language Models: A Large Organization Comparative Study