ROMCIR 2025: Overview of the 5th Workshop on Reducing Online Misinformation Through Credible Information Retrieval

The digital ecosystem is rife with various forms of information pollution that threaten individuals and society. Recent events, like the riots in England and Wales, allegedly incited by deliberate misinformation, highlight the pervasive impact of false news. Beyond riots, misinformation influences public opinion on politics and finance, tarnishes or enhances business reputations through deceptive reviews, and misguides individuals with unverified medical advice. Amidst this landscape, ensuring access to accurate, relevant information is imperative to maintain an unbiased perception of reality. This urgency has spurred interest in strategies to combat misinformation across diverse contexts and tasks. The ROMCIR Workshop addresses these challenges, engaging the Information Retrieval (IR) community to develop solutions beyond traditional misinformation detection. Its key objectives include identifying factors influencing credibility and truthfulness and integrating these as dimensions of relevance within Information Retrieval Systems (IRSs). The Workshop also focuses on early misinformation detection, ensuring retrieved search results are accurate and explainable. Furthermore, it emphasizes understanding the dual role of generative models, such as Large Language Models (LLMs), in exacerbating and mitigating misinformation, as well as the potential of human-in-the-loop systems to enhance IR accuracy and reliability.