Yuqi Hu

Mr. Yuqi Hu

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 February 2026 in Research Square
Verified
Preprint
Open access
PDF file
YHYuqi Hu
Yuqi Hu
YWYujing Wang
Yujing Wang
BWBrandon Watanabe
Brandon Watanabe
MGMingkai Gao
Mingkai Gao
ZQZhenzhen Qin
Zhenzhen Qin
ABAkram Bayat
Akram Bayat
Speech-language pathology (SLP) training programs face persistent challenges in providing sufficient opportunities for clinical placements or simulations that allow students to practice clinical communication with patients with communication disorders. We designed and implemented V.O.I.C.E., an AI-powered virtual patient system that enables realistic, open-ended clinical simulation for SLP students. V.O.I.C.E. integrates (1) a large language model (LLM) role-play agent designed to produce responses consistent with symptoms of post-stroke expressive aphasia (Broca’s aphasia), (2) a multimodal emotional expression pipeline that renders context-appropriate emotions of the virtual patient through coordinated speech prosody, facial expressions, and gestures, and (3) an LLM-based, rubric-guided debriefing module that generates structured formative feedback across core clinical communication competencies. We conducted a pilot mixed-methods, quasi-experimental evaluation with 11 graduate-level SLP students, where they completed three simulation sessions using the V.O.I.C.E system with progressive challenges. Results show significant gains in students' overall self-efficacy for interacting with Broca's aphasia patients in clinical settings. The self-efficacy gains were moderated by students’ perceptions of multimodal realism of the simulations. Qualitative findings highlight the system’s value as a low-stakes “bridge” for transition to clinical practice, which supports learning through iterative practice, reflection, and strategy refinement. This work provides a framework for AI-based simulation training applications that emphasizes interpersonal communication.

25 March 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
YHYuqi Hu
Yuqi Hu
Cannabis use remains a significant public health concern, yet access to evidence-based treatment is limited. This paper proposes a hybrid artificial intelligence chatbot framework for delivering psychosocial interventions for cannabis use. The model combines rule-based components to maintain structure, safety, and intervention fidelity with generative AI to enable more natural, personalized, and responsive interactions. By integrating these approaches, hybrid chatbots may expand access to scalable, private, and on-demand support while addressing key limitations of fully scripted or fully generative systems. This framework highlights the potential of AI-assisted intervention delivery and the need for future research on safety, effectiveness, and implementation.

26 October 2026 in International Journal of Public Health
Verified
Journal Article
Open access
PDF file
YHYuqi Hu
Yuqi Hu
Cannabis use and dependence is a rising public health challenge worldwide. Globally, the number of people with cannabis use disorder and its associated burden of diseases have notably increased in the past decades, particularly affecting young people [1,2]. Despite this rising burden, public health responses remain limited.While there are no established pharmacological treatments for cannabis use and dependence, psychosocial approaches, including motivational and cognitive-behavioral interventions, may improve cannabis use outcomes [3]. However, access to these treatments remains limited by workforce, time and transportation barriers. Treatment-related stigma, shaped by histories of drug control and enforcement that unevenly affect underserved populations, also discourages formal help-seeking [4]. Moreover, existing interventions are rarely delivered when people need support most and are often minimally responsive to changing risk states and individual needs. In this context, recent advances in artificial intelligence (AI) create an opportunity to address these gaps by making cannabis interventions more accessible, scalable, timely, and personalized.AI conversational chatbots are one promising approach. In practice, chatbots could be engineered to deliver core elements of psychosocial interventions for cannabis use by helping users set goals, identify triggers, strengthen motivation, and practice coping and relapse-prevention skills. A recent systematic review of chatbot-assisted substance use interventions, mainly for tobacco and alcohol use, found that experimental studies reported meaningful effects on substance use outcomes and users generally found the interventions useful and acceptable [5]. Importantly, chatbots could address existing access gaps by providing private, on-demand support outside conventional clinic settings and extending services to people facing geographic, transportation, scheduling, or stigma-related barriers. They could scale up at exceptionally low marginal costs, reducing reliance on scarce trained professionals. AI could also improve the timeliness of cannabis use interventions through just-in-time adaptive interventions (JITAIs). JITAIs can use machine learning methods to identify moments of heightened vulnerability and trigger support in a timely manner, although evidence of their effectiveness for substance use remains limited [6]. For cannabis use, this could involve drawing on self-reported data, prior engagement patterns, time, location, or other digital signals to identify vulnerable moments, such as craving after work, social exposure to use, boredom at night, or stress after conflict. The intervention could then provide a real-time response, such as a motivational message, coping prompt, or brief cognitive-behavioral exercise. Because cannabis use risk is often episodic and context-dependent, JITAIs could complement psychosocial approaches by delivering their content at high-risk moments rather than only in scheduled or user-initiated sessions.Conversational chatbots and JITAIs could be combined in intervention design. In such a model, the JITAI component identifies moments of heightened risk using self-reported, behavioral, or contextual data, while the chatbot delivers support in a conversational format, which could be personalized to adapt to the user's current state. For example, a recent smoking-cessation intervention integrated JITAI with a generative chatbot and was evaluated favorably by a domain expert [7].While direct evidence remains limited, certain elements of psychosocial interventions for tobacco use may be adaptable to cannabis interventions, as these substances overlap in psychosocial processes such as motivation, cravings, triggers, coping, and relapse. This possibility is supported by evidence that cannabis co-users who reduced cigarette use in a smoking cessation intervention may concurrently reduce their cannabis use, even though cannabis use was not explicitly addressed [8]. Therefore, existing evidence on tobacco interventions [5,7] could support the feasibility of AI-enabled cannabis use interventions.Nevertheless, adaptation would be required to account for the unique characteristics of cannabis use, including diverse products and routes of administration, self-medication, associations with social identity, and legal status. Above all, innovative AI interventions for cannabis use must be designed with safety and privacy as core requirements. A review found over half of health-related chatbots failed to respond adequately to high-risk messages such as suicidal thoughts [9]. Such limitations may become particularly consequential in cannabis interventions, where users may experience psychotic symptoms or other crises. AI systems also raise privacy concerns because they may collect sensitive behavioral data, infer vulnerability, and transmit information across systems that users do not fully understand [10]. Depending on the jurisdiction, intervention data may be subject to compelled disclosure in legal or administrative proceedings. Systems should therefore have a clearly limited clinical scope, minimize data collection and retention, clearly explain any limits of confidentiality, and provide reliable escalation pathways to human support.In summary, AI approaches could substantially extend the reach of psychosocial support for cannabis use and warrant more sustained research. Design, development and evaluation of innovative AI interventions should extend beyond a focus on efficacy to considerations of whether these tools improve engagement and reduce access and stigma-related barriers without safety or privacy concerns. Future implementation research should identify which populations benefit or remain underserved, establish reliable pathways to human care, and examine how effective innovations can be integrated into routine care and public health infrastructure.

14 July 2026 in JMIR Preprints
Verified
Preprint
Open access
YHYuqi Hu
Yuqi Hu
YWYujing Wang
Yujing Wang
BWBrandon Watanabe
Brandon Watanabe
MGMingkai Gao
Mingkai Gao
ZQZhenzhen Qin
Zhenzhen Qin
ABAkram Bayat
Akram Bayat
BACKGROUND Speech-language pathology (SLP) students need repeated opportunities to practice clinical communication with patients who have communication disorders, including aphasia. Existing simulations support case-based learning, but many provide limited open-ended dialogue and limited affective feedback during communication breakdowns. OBJECTIVE This study aimed to describe and pilot-evaluate Virtual Observation and Interactive Clinical Education (V.O.I.C.E.), an emotionally responsive artificial intelligence (AI) virtual patient system for aphasia-focused communication training in graduate SLP education. METHODS We conducted a single-group, mixed methods pilot study with 11 graduate SLP students. V.O.I.C.E. combines a large language model patient persona with poststroke Broca aphasia, a multimodal affect pipeline that renders patient emotion through speech prosody, facial expression, and gesture, and an automated rubric-guided debriefing report. Participants completed three approximately 10-minute simulations with progressively more severe aphasia presentations. Quantitative measures included pre-post self-efficacy for clinical communication and poststudy ratings of dialogue realism, voice realism, facial/gesture realism, usability, and educational value, which were analyzed descriptively. Qualitative measures including open-ended questions examining participants’ experience with and attitudes towards the AI patient, which were analyzed using hybrid deductive-inductive content analysis. RESULTS Overall self-efficacy improved significantly after the simulations, with a mean within-person gain of 1.09 points on a 6-point scale (95% CI 0.33-1.84; p=.03). Significant gains were observed for confidence in interviewing clients with aphasia (mean 2.33 to 3.89; change +1.56; p=.04) and responding to unexpected patient behavior (mean 3.73 to 5.09; change +1.36; p=.03). Confidence gains were positively associated with perceived dialogue realism (Spearman rho=0.79; p=.01), facial/gesture realism (rho=0.81; p=.01), and voice realism (rho=0.75; p=.02). Qualitatively, students described V.O.I.C.E. as a safe first step toward clinical practice that supported iterative practice, feedback-driven strategy refinement, metacognitive awareness, and rehearsal of emotionally challenging interactions; they also identified limitations related to speech recognition, emotional subtlety, feedback specificity, and the need for direct patient contact. CONCLUSIONS This pilot study suggests that emotionally responsive AI virtual patients may provide a useful opportunity for SLP students in their early stages of learning and practicing aphasia-focused clinical communication. The findings highlight the importance of multimodal realism, emotionally contingent patient behavior, and actionable debriefing. Larger controlled studies are needed to further evaluate the efficacy of the V.O.I.C.E. system.

25 August 2026 in Frontiers in Communication
Verified
Journal Article
Open access
PDF file
YHYuqi Hu
Yuqi Hu
Health misinformation on social media is not only a problem of inaccurate content, but also a communication problem shaped by trust, uncertainty, emotion, identity, platform design, and unequal access to credible information. Artificial intelligence (AI) is increasingly proposed as a tool for addressing this problem, yet much of the literature emphasizes detection accuracy, automated moderation, or technical performance. This mini review reframes AI as part of a broader health communication response cycle for addressing health misinformation on social media. Drawing on recent literature, it examines how AI can support social listening, intervention design and delivery, and how these systems are evaluated. AI-enabled social listening can help identify emerging misinformation narratives, information voids, audience concerns, emotional responses, and diffusion patterns. These insights may inform corrective messages, prebunking content, plain-language explanations, multilingual communication, chatbot-based engagement, and audience-tailored responses. It should be noted that the current evidence base remains concentrated in COVID-19 and vaccine-related contexts, with more limited evidence from other health domains. Moreover, existing studies often evaluate technical performance rather than real-world communication outcomes, and evidence for AI-enabled interventions remains mixed and insufficient. Overall, AI has potential to support health communication against misinformation on social media, but it should not replace health communicators, public health professionals, clinicians, or trusted community messengers. Instead, its value depends on transparent, human-in-the-loop, theory-informed, and community-engaged implementation.