Vol. 3 No. 1 (2025)

This issue of the Journal of Online Trust and Safety, which serves as the Conference Proceedings for the 2025 Trust and Safety Research Conference, consists of two research articles and three commentaries contributed by legal scholars, academic researchers, and industry practitioners.

In the first research article, Dahlke & Hancock examine digital trace data and use a double machine learning approach to disentangle the relationship between exposure to untrustworthy websites and belief in false election claims, finding asymmetric effects that were strongest among conservatives and increased with additional exposures. In the second, Xu et al. conduct in-depth interviews with young adults in India and the US, develop the concept of "information modes" to describe the temporary states that can alter how people seek out and evaluate content, and emphasize the need for digital literacy approaches that take the social and emotional dimensions of online engagement into account. In the first commentary, Li et al. propose a hybrid model for Community Notes where Large Language Models (LLMs) can increase the speed and scale of note writing, humans retain the final decision on helpfulness ratings, and the data collected from both can create a mutually beneficial feedback cycle that the authors term Reinforcement Learning from Community Feedback (RLCF). In the second, Radeksy & Hiniker examine Problematic Media Use (PMU) and its behavioral correlates as a content-agnostic approach for measuring harms related to the extreme overuse of online platforms, which could potentially help in mitigating those harms, and examine the value of such measures for policymakers, academics, and industry practitioners.  In the third, Gerbrandt & Howard discuss the "newsworthiness" exception in online platforms' content moderation policies, with a focus on cases from Facebook, and argue that such ad hoc exemptions, especially from politicians, could be replaced with more nuanced rules that consider the inferred purpose of the user's post.

The Journal of Online Trust and Safety is grateful to the Omidyar Network for their generous support, and our reviewers for their timely and constructive input.