Like a Hammer, It Can Build, It Can Break: Large Language Model Uses, Perceptions, and Adoption in Cybersecurity Operations on Reddit
A new research paper analyzes discussions on Reddit from December 2022 to September 2025 to understand how cybersecurity practitioners use and perceive large language models (LLMs). The study found that while LLMs offer efficiency gains for low-risk tasks, issues with reliability, verification, and security risks limit their autonomy in security operations. The research provides recommendations for developing and adopting LLM tools to enhance organizational security and practitioner safety. AI