Children with Autism Spectrum Disorder (ASD) often face challenges with musical engagement due to unique sensory and neural processing needs.To address this, we introduce uCue, a musical interface designed to enhance active musical engagement.This interactive playback system offers modular arrangements of children's songs at an accessible tempo, enabling listeners to manipulate musical layers and create personalized renditions in real-time.Deployed in listening sessions with seven parent-child dyads, uCue facilitated self-expression through singing and gestures while fostering emotional regulation and sensory engagement.Participants quickly adapted to the interface, preferred soothing sounds, and expressed interest in more rhythmic layers over time.Our findings suggest that uCue has the potential to enhance musical interactions by allowing children to explore and control auditory experiences.We also discuss how uCue and data from its logs might support therapeutic goals in music therapy for children with ASD and provide design recommendations for similar technologies.

Large Language Models (LLMs) can analyze large datasets semantically. However, research on applying LLMs for mental health text classification is relatively new and developing. Existing methods often use supervised, deep, and reinforcement learning, which rely heavily on fine-tuning and reward models. To investigate whether LLMs can assist in recommending mental health apps based on user reviews, our study collected approximately 200k user reviews from 73 mental health mobile applications. We instructed selected LLMs to classify individual reviews into 1-5 star ratings, subsequently averaging these results to derive an overall rating for each app reflecting current user feedback. While the best supervised learning method in our experiments achieved an F1-Score of 0.79 which required significantly more human effort, the GPT-4 and Gemini 1.5 Pro delivered a strong ‘out-of-the-box’ performance with an overall F1-Score of 0.76. We provide further statistical comparisons and discussions of the performance of these models for the text classification task. Using a crowdsourcing platform to determine agreement levels, we observed that human ratings align closely with GPT ratings. In addition, we analyze specific features and concerns highlighted in mental health app reviews. Alongside our analysis, we make our data available for further experimentation and benchmarking.

Srija Mukhopadhyay, Abhishek Rajgaria, Prerana Khatiwada, Manish Shrivastava, Dan Roth, Vivek Gupta. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Traditional history education often fails to cultivate historical empathy due to rigid curricula and limited opportunities for personalized, emotionally resonant engagement. We explore the potential of LLM-based historical personas to address these gaps by enabling students to engage in real-time, conversational interactions with simulated historical figures. A formative study with teachers and students surfaced key challenges and expectations around AI-mediated historical dialogue, informing the development of Baseline and Experimental HistoChat, AI persona systems featuring differing prompting strategies. A subsequent user study showed that these interactions fostered deeper inquiry, curiosity, and emotional engagement-while also revealing key limitations. From a CSCW perspective, this work expands the role of AI from task assistant to epistemic partner, contributing to ongoing discourse on how dialogic systems can support meaning-making, empathy, and co-constructed learning in educational settings. Our findings yield valuable insights into the impact of tailored AI interactions on personalized and empathetic history education.

Social media platforms have transformed global communication and interaction, with TikTok emerging as a critical tool for education, connection, and social impact, including in contexts where infrastructural resources are limited. Amid growing political discussions about banning platforms like TikTok, such actions can create significant ripple effects, particularly impacting marginalized communities. We present a study on Nepal, where a TikTok ban was recently imposed and lifted. As a low-resource country in transition where digital communication is rapidly evolving, TikTok enables a space for community engagement and cultural expression. In this context, we conducted an online survey ( N=108 ) to explore user values, experiences, and strategies for navigating online spaces post-ban. By examining these transitions, we aim to improve our understanding of how digital technologies, policy responses, and cultural dynamics interact globally and their implications for governance and societal norms. Our results indicate that users express skepticism toward platform bans but often passively accept them without active opposition. Findings suggest the importance of institutionalizing collective governance models that encourage public deliberation, nuanced control, and socially resonant policy decisions.

Misinformation permeates online news media, making it hard for users to trust and verify content. Building upon prior work that highlights online news challenges and the importance of digital literacy skills, we examine user digital media skills, online news consumption behaviors, and perceptions of current misinformation tools. To better understand these dynamics, we conducted a formative, mixed methods study (n=34) that included a survey, two weeks of browser-based application logging using a Chrome plugin, and follow-up semi-structured interviews. Contradictions in the survey and log results indicate that participants in our sample often overestimate their news consumption habits. While information and news media literacy scores are generally high, less than half (47%, 16/34) exhibit lateral reading. We provide insights into users' challenges in navigating today's information landscape and propose effective integrated solutions. Interview findings inform the design of online news interventions and personal informatics tools to further improve media literacy while maintaining user privacy.