There are growing concerns about the risks posed by AI companion applications designed for emotional engagement.Existing safety evaluations often rely on self-reported user data or interviews, offering limited insights into real-time dynamics.We present the first end-to-end scalable framework for controlled simulation and safety evaluation of multi-turn interactions with AI companion applications.Our framework integrates four key components: persona construction with clinical and psychometric validation, persona-specific scenario generation, scenario-driven multi-turn simulation with a dialogue refinement module that preserves persona fidelity, and harm evaluation.We apply this framework to evaluate how Replika, a widely used AI companion app, responds to high-risk user groups.We construct 9 personas representing individuals with depression, anxiety, PTSD, eating disorders, and incel identity, and collect 1,674 dialogue pairs across 25 high-risk scenarios.We combine emotion modeling and LLM-assisted utterance-and harm-level classification to analyze these exchanges.Results show that Replika exhibits a narrow emotional range dominated by curiosity and care, while frequently mirroring or normalizing unsafe content such as self-harm, disordered eating, and violentfantasy narratives.These findings highlight how controlled persona simulations can serve as a scalable testbed for evaluating safety risks in AI companions. 1 Content Warning: This paper includes examples of dialogues involving self-harm, disordered eating, and misogynistic language.

Despite being an integral tool for finding health-related information online, YouTube has faced criticism for disseminating COVID-19 misinformation globally to its users. Yet, prior audit studies have predominantly investigated YouTube within the Global North contexts, often overlooking the Global South. To address this gap, we conducted a comprehensive 10-day geolocation-based audit on YouTube to compare the prevalence of COVID-19 misinformation in search results between the United States (US) and South Africa (SA), the countries heavily affected by the pandemic in the Global North and the Global South, respectively. For each country, we selected 3 geolocations and placed sock-puppets, or bots emulating "real" users, that collected search results for 48 search queries sorted by 4 search filters for 10 days, yielding a dataset of 915K results. We found that 31.55% of the top-10 search results contained COVID-19 misinformation. Among the top-10 search results, bots in SA faced significantly more misinformative search results than their US counterparts. Overall, our study highlights the contrasting algorithmic behaviors of YouTube search between two countries, underscoring the need for the platform to regulate algorithmic behavior consistently across different regions of the Globe. Warning: We caution the readers that some examples provided to better contextualize our data can be offensive.

As more users turn to video-sharing platforms like YouTube as an information source, they may consume misinformation despite their best efforts. In this work, we investigate ways that users can better assess the credibility of videos by first exploring how users currently determine credibility using existing signals on platforms and then by introducing and evaluating new credibility-based signals. We conducted 12 contextual inquiry interviews with YouTube users, determining that participants used a combination of existing signals, such as the channel name, the production quality, and prior knowledge, to evaluate credibility, yet sometimes stumbled in their efforts to do so. We then developed Viblio, a prototype system that enables YouTube users to view and add citations and related information while watching a video based on our participants’ needs. From an evaluation with 12 people, all participants found Viblio to be intuitive and useful in the process of evaluating a video’s credibility and could see themselves using Viblio in the future.

There is a growing demand for transparency in search engines to understand how search results are curated and to enhance users’ trust. Prior research has introduced search result explanations with a focus on how to explain, assuming explanations are beneficial. Our study takes a step back to examine if search explanations are needed and when they are likely to provide benefits. Additionally, we summarize key characteristics of helpful explanations and share users’ perspectives on explanation features provided by Google and Bing. Interviews with non-technical individuals reveal that users do not always seek or understand search explanations and mostly desire them for complex and critical tasks. They find Google’s search explanations too obvious but appreciate the ability to contest search results. Based on our findings, we offer design recommendations for search engines and explanations to help users better evaluate search results and enhance their search experience.

With the 2022 US midterm elections approaching, conspiratorial claims about the 2020 presidential elections continue to threaten users’ trust in the electoral process. To regulate election misinformation, YouTube introduced policies to remove such content from its searches and recommendations. In this paper, we conduct a 9-day crowd-sourced audit on YouTube to assess the extent of enactment of such policies. We recruited 99 users who installed a browser extension that enabled us to collect up-next recommendation trails and search results for 45 videos and 88 search queries about the 2020 elections. We find that YouTube’s search results, irrespective of search query bias, contain more videos that oppose rather than support election misinformation. However, watching misinformative election videos still lead users to a small number of misinformative videos in the up-next trails. Our results imply that while YouTube largely seems successful in regulating election misinformation, there is still room for improvement.

Increasing demands for fact-checking have led to a growing interest in developing systems and tools to automate the fact-checking process. However, such systems are limited in practice because their system design often does not take into account how fact-checking is done in the real world and ignores the insights and needs of various stakeholder groups core to the fact-checking process. This paper unpacks the fact-checking process by revealing the infrastructures---both human and technological---that support and shape fact-checking work. We interviewed 26 participants belonging to 16 fact-checking teams and organizations with representation from 4 continents. Through these interviews, we describe the human infrastructure of fact-checking by identifying and presenting, in-depth, the roles of six primary stakeholder groups, 1) Editors, 2) External fact-checkers, 3) In-house fact-checkers, 4) Investigators and researchers, 5) Social media managers, and 6) Advocators. Our findings highlight that the fact-checking process is a collaborative effort among various stakeholder groups and associated technological and informational infrastructures. By rendering visibility to the infrastructures, we reveal how fact-checking has evolved to include both short-term claims centric and long-term advocacy centric fact-checking. Our work also identifies key social and technical needs and challenges faced by each stakeholder group. Based on our findings, we suggest that improving the quality of fact-checking requires systematic changes in the civic, informational, and technological contexts.

There is a growing concern that e-commerce platforms are amplifying vaccine-misinformation. To investigate, we conduct two-sets of algorithmic audits for vaccine misinformation on the search and recommendation algorithms of Amazon—world’s leading e-retailer. First, we systematically audit search-results belonging to vaccine-related search-queries without logging into the platform—unpersonalized audits. We find 10.47% of search-results promote misinformative health products. We also observe ranking-bias, with Amazon ranking misinformative search-results higher than debunking search-results. Next, we analyze the effects of personalization due to account-history, where history is built progressively by performing various real-world user-actions, such as clicking a product. We find evidence of filter-bubble effect in Amazon’s recommendations; accounts performing actions on misinformative products are presented with more misinformation compared to accounts performing actions on neutral and debunking products. Interestingly, once user clicks on a misinformative product, homepage recommendations become more contaminated compared to when user shows an intention to buy that product.

Search engines are the primary gateways of information. Yet, they do not take into account the credibility of search results. There is a growing concern that YouTube, the second largest search engine and the most popular video-sharing platform, has been promoting and recommending misinformative content for certain search topics. In this study, we audit YouTube to verify those claims. Our audit experiments investigate whether personalization (based on age, gender, geolocation, or watch history) contributes to amplifying misinformation. After shortlisting five popular topics known to contain misinformative content and compiling associated search queries representing them, we conduct two sets of audits-Search-and Watch-misinformative audits. Our audits resulted in a dataset of more than 56K videos compiled to link stance (whether promoting misinformation or not) with the personalization attribute audited. Our videos correspond to three major YouTube components: search results, Up-Next, and Top 5 recommendations. We find that demographics, such as, gender, age, and geolocation do not have a significant effect on amplifying misinformation in returned search results for users with brand new accounts. On the other hand, once a user develops a watch history, these attributes do affect the extent of misinformation recommended to them. Further analyses reveal a filter bubble effect, both in the Top 5 and Up-Next recommendations for all topics, except vaccine controversies; for these topics, watching videos that promote misinformation leads to more misinformative video recommendations. In conclusion, YouTube still has a long way to go to mitigate misinformation on its platform.

Transparency in moderation practices is crucial to the success of an online community. To meet the growing demands of transparency and accountability, several academics came together and proposed the Santa Clara Principles on Transparency and Accountability in Content Moderation (SCP). In 2018, Reddit, home to uniquely moderated communities called subreddits, announced in its transparency report that the company is aligning its content moderation practices with the SCP. But do the moderators of subreddit communities follow these guidelines too? In this paper, we answer this question by employing a mixed-methods approach on public moderation logs collected from 204 subreddits over a period of five months, containing more than 0.5M instances of removals by both human moderators and AutoModerator. Our results reveal a lack of transparency in moderation practices. We find that while subreddits often rely on AutoModerator to sanction newcomer posts based on karma requirements and moderate uncivil content based on automated keyword lists, users are neither notified of these sanctions, nor are these practices formally stated in any of the subreddits' rules. We interviewed 13 Reddit moderators to hear their views on different facets of transparency and to determine why a lack of transparency is a widespread phenomenon. The interviews reveal that moderators' stance on transparency is divided, there is a lack of standardized process to appeal against content removal and Reddit's app and platform design often impede moderators' ability to be transparent in their moderation practices.

Issue Tracking Systems (ITS) such as Bugzilla can be viewed as Process Aware Information Systems (PAIS) generating event-logs during the life-cycle of a bug report. Process Mining consists of mining event logs generated from PAIS for process model discovery, conformance and enhancement. We apply process map discovery techniques to mine event trace data generated from ITS of open source Firefox browser project to generate and study process models. Bug life-cycle consists of diversity and variance. Therefore, the process models generated from the event-logs are spaghetti-like with large number of edges, inter-connections and nodes. Such models are complex to analyse and difficult to comprehend by a process analyst. We improve the Goodness (fitness and structural complexity) of the process models by splitting the event-log into homogeneous subsets by clustering structurally similar traces. We adapt the K-Medoid clustering algorithm with two different distance metrics: Longest Common Subsequence (LCS) and Dynamic Time Warping (DTW). We evaluate the goodness of the process models generated from the clusters using complexity and fitness metrics. We study back-forth and self-loops, bug reopening, and bottleneck in the clusters obtained and show that clustering enables better analysis. We also propose an algorithm to automate the clustering process-the algorithm takes as input the event log and returns the best cluster set.