This research explores how large language models can effectively evaluate online news publisher reliability at scale, reducing dependence on limited human expert resources.

• Tested LLM capability to assess expert-designed reliability criteria in Italian news media
• Compared model evaluations with human expert assessments
• Demonstrated potential for AI-assisted news credibility frameworks
• Addressed information security challenges related to misinformation detection

This work offers significant implications for fighting misinformation and establishing automated trust metrics in digital media ecosystems, providing a scalable approach to online content verification critical for information security.

Evaluation of Reliability Criteria for News Publishers with Large Language Models