On the Effectiveness of Fact Checking Information from Politically Congruent and Incongruent Large Language Models

Collection of recent AI/CS papers focusing on large language model (LLM) deployment, governance, and datasets. Key findings and risks: (1) LLM-based fact-checkers can shift user trust in true and false political news even when politically incongruent, but they also risk tainting truth at scale when wrong or inconclusive; (2) EHR audit logs framed as multi-axial event streams highlight representation, privacy, and governance needs for learning and monitoring clinical workflows; (3) Complete Trip is a linked multimodal human mobility dataset reconstructed from smartphone LBS data, enabling route