Early-stage interdisciplinary research ideation is often challenged by limited expert access, uncertainty about what to ask, and the cognitive burden of synthesizing unfamiliar domain perspectives. This paper presents Perspectra, a forum-style multi-agent system that structures and visualizes deliberation among LLM-simulated domain experts to support exploration and refinement of emerging research ideas, while encouraging critical thinking and reflections. The interface design combines 1) a threaded canvas for parallel topic exploration with visualization of agent discourse dynamics informed by argumentation theory to aid sensemaking; and 2) feature that enables users to invite multiple self-chosen agents into an ongoing discussion. We conducted a user study with 18 participants, comparing Perspectra against a vanilla chat baseline given a task for the user to develop a short research proposal. Our findings show that Perspectra’s design elicits significantly more higher-order critical thinking behaviors during interactions with agents when compared to a traditional chat interface. We also observed more interdisciplinary user replies via forum-styled design, and more frequent and structured proposal revisions (rather than unstructured note-taking). Based on our findings, we further contribute interaction design implications of using multi-agent deliberation for complex ideation and knowledge search, combining flexibility with structured exploration to support user sensemaking and critical thinking.

○ where users can indicate their topics of interest for exploration.Persona Nodes 2 ○ represent AI-simulated expert perspectives that can suggest related literature retrieved from online publication database (Literature Nodes 3 ○), and subsequently provide feedback and critiques (Critique Nodes 4 ○) to users' initial research idea.Based on the critiques and identified literature, the system can further help revise users' initial idea into a revised RQ (RQ node 5 ○).Users can perform this process iteratively and combine inputs from multiple expert personas until they discover satisfactory RQs of their interest.

Large Language Models (LLMs) are increasingly used by teenagers and young adults in everyday life, ranging from emotional support and creative expression to educational assistance. However, their unique vulnerabilities and risk profiles remain under-examined in current safety benchmarks and moderation systems, leaving this population disproportionately exposed to harm. In this work, we present Youth AI Risk (YAIR), the first benchmark dataset designed to evaluate and improve the safety of youth–LLM interactions. YAIR consists of 12,449 annotated conversation snippets spanning 78 fine-grained risk types, grounded in a taxonomy of youth-specific harms such as grooming, boundary violation, identity confusion, and emotional overreliance. We systematically evaluate widely adopted moderation models on YAIR and find that existing approaches substantially underperform in detecting youth-centered risks, often missing contextually subtle yet developmentally harmful interactions. To address these gaps, we introduce YouthSafe, a real-time risk detection model optimized for youth–GenAI contexts. YouthSafe significantly outperforms prior systems across multiple metrics on risk detection and classification, offering a concrete step toward safer and more developmentally appropriate AI interactions for young users.

Emotional support is a crucial aspect of communication between community members and police dispatchers during incident reporting. However, there is a lack of understanding about how emotional support is delivered through text-based systems, especially in various non-emergency contexts. In this study, we analyzed two years of chat logs comprising 57,114 messages across 8,239 incidents from 130 higher education institutions. Our empirical findings revealed significant variations in emotional support provided by dispatchers, influenced by the type of incident, service time, and a noticeable decline in support over time across multiple organizations. To improve the consistency and quality of emotional support, we developed and implemented a fine-tuned Large Language Model (LLM), named dispatcherLLM, designed to suggest replies through simulating human dispatchers' languages with appropriate emotional support. We evaluated dispatcherLLM by comparing its generated responses to those of human dispatchers and other off-the-shelf models using real chat messages. Additionally, we conducted a human evaluation to assess the perceived effectiveness of the support provided by dispatcherLLM. This study not only contributes new empirical understandings of emotional support in text-based dispatch systems but also demonstrates the significant potential of generative AI in improving service delivery.

Developing novel research questions (RQs) often requires extensive literature reviews, especially in interdisciplinary fields. To support RQ development through human-AI co-creation, we leveraged Large Language Models (LLMs) to build an LLM-based agent system named CoQuest. We conducted an experiment with 20 HCI researchers to examine the impact of two interaction designs: breadth-first and depth-first RQ generation. The findings revealed that participants perceived the breadth-first approach as more creative and trustworthy upon task completion. Conversely, during the task, participants considered the depth-first generated RQs as more creative. Additionally, we discovered that AI processing delays allowed users to reflect on multiple RQs simultaneously, leading to a higher quantity of generated RQs and an enhanced sense of control. Our work makes both theoretical and practical contributions by proposing and evaluating a mental model for human-AI co-creation of RQs. We also address potential ethical issues, such as biases and over-reliance on AI, advocating for using the system to improve human research creativity rather than automating scientific inquiry. The system’s source is available at: https://github.com/yiren-liu/coquest.

Recently, an increasing number of safety organizations in the U.S. have incorporated text-based risk reporting systems to respond to safety incident reports from their community members. To gain a better understanding of the interaction between community members and dispatchers using text-based risk reporting systems, this study conducts a system log analysis ofLiveSafe, a community safety reporting system, to provide empirical evidence of the conversational patterns between users and dispatchers using both quantitative and qualitative methods. We created an ontology to capture information (e.g., location, attacker, target, weapon, start-time, and end-time, etc.) that dispatchers often collected from users regarding their incident tips. Applying the proposed ontology, we found that dispatchers often asked users for different information across varied event types (e.g.,Attacker forAbuse andAttack events,Target forHarassment events). Additionally, using emotion detection and regression analysis, we found an inconsistency in dispatchers' emotional support and responsiveness to users' messages between different organizations and between incident categories. The results also showed that users had a higher response rate and responded quicker when dispatchers provided emotional support. These novel findings brought significant insights to both practitioners and system designers, e.g., AI-based solutions to augment human agents' skills for improved service quality.

Industrial data are usually collinear, which can cause pure data-driven sparse learning to deselect physically relevant variables and select collinear surrogates. In this article, a novel two-step learning approach to retaining knowledge-informed variables (KIVs) is proposed to build inferential models. The first step is an improved knowledge-informed Lasso (KILasso) algorithm by removing penalty on the KIVs to produce a series of candidate subsets that guarantee the retention of the KIVs. The candidate subsets are then used to run the KILasso or ridge regression again to select the best sets of variables and estimate the final model. Two new algorithms are proposed and applied to datasets from an industrial boiler process and the Dow Chemical challenge problem. It is demonstrated that some important physically relevant variables are deselected by pure data-driven sparse methods, but they are retained using the proposed knowledge-informed methods with superior prediction performance.

Early conversational agents (CAs) focused on dyadic human-AI interaction between humans and the CAs, followed by the increasing popularity of polyadic human-AI interaction, in which CAs are designed to mediate human-human interactions. CAs for polyadic interactions are unique because they encompass hybrid social interactions, i.e., human-CA, human-to-human, and human-to-group behaviors. However, research on polyadic CAs is scattered across different fields, making it challenging to identify, compare, and accumulate existing knowledge. To promote the future design of CA systems, we conducted a literature review of ACM publications and identified a set of works that conducted UX (user experience) research. We qualitatively synthesized the effects of polyadic CAs into four aspects of human-human interactions, i.e., communication, engagement, connection, and relationship maintenance. Through a mixed-method analysis of the selected polyadic and dyadic CA studies, we developed a suite of evaluation measurements on the effects. Our findings show that designing with social boundaries, such as privacy, disclosure, and identification, is crucial for ethical polyadic CAs. Future research should also advance usability testing methods and trust-building guidelines for conversational AI.