Large Language Models in Cybersecurity: A PRISMA-ScR Guided Scoping Review of Threats, Defenses, and Emerging Applications

Keyphrases

• Cybersecurity 100%
• PRISMA-ScR 100%
• Large Language Models 100%
• Scoping Review 100%
• Threat Defense 100%
• Exclusion Criteria 50%
• Offensive 50%
• Scopus 25%
• Evidence-based 25%
• Hallucinations 25%
• Web of Science 25%
• Evolving Roles 25%
• General Purpose Systems 25%
• Narrative Synthesis 25%
• Generation Capacity 25%
• Governance Challenges 25%
• Threat Detection 25%
• January 2017 25%
• Natural Language Processing 25%
• Data Standardization 25%
• Model Reliability 25%
• Defensive Response 25%
• Incident Response 25%
• Code Generation 25%
• Technical Perspective 25%
• Evaluation Standard 25%
• Adversarial Attack 25%
• IEEE Xplore 25%
• ACM Digital Library 25%
• Dual Use 25%
• Vulnerability Detection 25%
• Operational Perspective 25%
• Deployment Framework 25%
• Model Governance 25%
• Security Analysts 25%
• Current Landscape 25%
• ArXiv 25%
• Evaluation Benchmark 25%
• Reasoning Capabilities 25%
• Large Language Model Application 25%
• Model Privacy 25%
• Security Automation 25%
• Code Reasoning 25%
• Phishing 25%
• Vulnerability Management 25%
• Malware Evolution 25%
• Governance Perspective 25%

Computer Science

• Cybersecurity 100%
• Large Language Model 100%
• Emerging Application 100%
• Natural Language Processing 20%
• Code Generation 20%
• Malware 20%
• Research Direction 20%
• Structured Search 20%
• Web of Science 20%
• Gray Literature 20%
• Adversarial Machine Learning 20%
• Incident Response 20%
• Threat Detection 20%
• Digital Library 20%
• Evaluation Benchmark 20%
• Vulnerability Management 20%
• Security Analyst 20%
• Security Automation 20%