This research explores how LLM-generated summaries of web content can improve the efficiency and accuracy of crowd-sourced fact-checking processes.

• Compared performance of human assessments using full web pages versus AI-generated summaries
• Demonstrated that condensed LLM summaries enable faster fact-checking without sacrificing quality
• Revealed that properly designed AI assistance can enhance human verification capabilities
• Established a cost-effective approach to combat online misinformation at scale

Security Impact: As mis/disinformation proliferates online, this approach provides a practical, scalable solution for information verification systems that can strengthen digital security and protect users from harmful falsehoods.

Collecting Cost-Effective, High-Quality Truthfulness Assessments with LLM Summarized Evidence