This research systematically evaluates five leading LLMs (ChatGPT-4, Llama 3, Llama 3.1, Claude 3.5, and Gemini) on their ability to fact-check political statements across a massive dataset of 16,500+ professional fact-checks.

Key Findings:

• Comparative performance analysis of major LLMs in detecting true vs. false political information
• Identifies strengths and limitations of AI-powered fact-checking across diverse political topics
• Establishes benchmarks for evaluating LLM reliability in misinformation detection
• Provides insights for developing more effective automated fact-checking systems

Security Implications: This research is crucial for understanding how LLMs can be deployed as front-line defenses against misinformation campaigns, helping organizations identify reliable AI tools for information verification and developing more robust approaches to combat digital threats.

Fact-checking with Generative AI: A Systematic Cross-Topic Examination of LLMs Capacity to Detect Veracity of Political Information