Mental Health Chatbot for Young Adults With Depressive Symptoms During the COVID-19 Pandemic: Single-Blind, Three-Arm Randomized Controlled Trial

Journal of Medical Internet Research, Journal Year: 2022, Volume and Issue: 24(11), P. e40719 - e40719

Published: Nov. 3, 2022

Depression has a high prevalence among young adults, especially during the COVID-19 pandemic. However, mental health services remain scarce and underutilized worldwide. Mental chatbots are novel digital technology to provide fully automated interventions for depressive symptoms.The purpose of this study was test clinical effectiveness nonclinical performance cognitive behavioral therapy (CBT)-based chatbot (XiaoE) adults with symptoms pandemic.In single-blind, 3-arm randomized controlled trial, participants manifesting recruited from Chinese university were randomly assigned (XiaoE; n=49), an e-book (n=49), or general (Xiaoai; n=50) group in ratio 1:1:1. Participants received 1-week intervention. The primary outcome reduction according 9-item Patient Health Questionnaire (PHQ-9) at 1 week later (T1) month (T2). Both intention-to-treat per-protocol analyses conducted under analysis covariance models adjusting baseline data. Controlled multiple imputation δ-based sensitivity performed missing secondary outcomes level working alliance measured using Working Alliance (WAQ), usability Usability Metric User Experience-LITE (UMUX-LITE), acceptability Acceptability Scale (AS).Participants on average 18.78 years old, 37.2% (55/148) female. mean PHQ-9 score 10.02 (SD 3.18; range 2-19). Intention-to-treat revealed lower scores XiaoE compared Xiaoai both T1 (F2,136=17.011; P<.001; d=0.51) T2 (F2,136=5.477; P=.005; d=0.31). Better (WAQ; F2,145=3.407; P=.04) (AS; F2,145=4.322; P=.02) discovered XiaoE, while no significant difference arms found (UMUX-LITE; F2,145=0.968; P=.38).A CBT-based is feasible engaging therapeutic approach that allows easy accessibility self-guided assistance symptoms. A systematic evaluation metrics been established study. In future, focus necessary explore mechanism by which work patients. Further evidence required confirm long-term via trails replicated longer dose, as well exploration its stronger efficacy comparison other active controls.Chinese Clinical Trial Registry ChiCTR2100052532; http://www.chictr.org.cn/showproj.aspx?proj=135744.

Language: Английский

---

Roles, Users, Benefits, and Limitations of Chatbots in Health Care: Rapid Review

Journal of Medical Internet Research, Journal Year: 2024, Volume and Issue: 26, P. e56930 - e56930

Published: April 12, 2024

Background Chatbots, or conversational agents, have emerged as significant tools in health care, driven by advancements artificial intelligence and digital technology. These programs are designed to simulate human conversations, addressing various care needs. However, no comprehensive synthesis of chatbots’ roles, users, benefits, limitations is available inform future research application the field. Objective This review aims describe characteristics, focusing on their diverse roles pathway, user groups, limitations. Methods A rapid published literature from 2017 2023 was performed with a search strategy developed collaboration sciences librarian implemented MEDLINE Embase databases. Primary studies reporting chatbot benefits were included. Two reviewers dual-screened results. Extracted data subjected content analysis. Results The categorized into 2 themes: delivery remote services, including patient support, management, education, skills building, behavior promotion, provision administrative assistance providers. User groups spanned across patients chronic conditions well cancer; individuals focused lifestyle improvements; demographic such women, families, older adults. Professionals students also alongside seeking mental behavioral change, educational enhancement. chatbots classified improvement quality efficiency cost-effectiveness delivery. identified encompassed ethical challenges, medicolegal safety concerns, technical difficulties, experience issues, societal economic impacts. Conclusions Health offer wide spectrum applications, potentially impacting aspects care. While they promising for improving quality, integration system must be approached consideration ensure optimal, safe, equitable use.

Language: Английский

---

CloChat: Understanding How People Customize, Interact, and Experience Personas in Large Language Models

Published: May 11, 2024

Large language models (LLMs) have facilitated significant strides in generating conversational agents, enabling seamless, contextually relevant dialogues across diverse topics. However, the existing LLM-driven agents fixed personalities and functionalities, limiting their adaptability to individual user needs. Creating personalized agent personas with distinct expertise or traits can address this issue. Nonetheless, we lack knowledge of how people customize interact personas. In research, investigated users impact on interaction quality, diversity, dynamics. To end, developed CloChat, an interface supporting easy accurate customization LLMs. We conducted a study comparing participants CloChat ChatGPT. The results indicate that formed emotional bonds customized engaged more dynamic dialogues, showed interest sustaining interactions. These findings contribute design implications for future systems using

Language: Английский

---

Waiting for a digital therapist: three challenges on the path to psychotherapy delivered by artificial intelligence

Frontiers in Psychiatry, Journal Year: 2023, Volume and Issue: 14

Published: June 1, 2023

Growing demand for broadly accessible mental health care, together with the rapid development of new technologies, trigger discussions about feasibility psychotherapeutic interventions based on interactions Conversational Artificial Intelligence (CAI). Many authors argue that while currently available CAI can be a useful supplement human-delivered psychotherapy, it is not yet capable delivering fully fledged psychotherapy its own. The goal this paper to investigate what are most important obstacles our way developing systems in future. To end, we formulate and discuss three challenges central quest. Firstly, might able develop effective AI-based unless deepen understanding makes effective. Secondly, assuming requires building therapeutic relationship, clear whether delivered by non-human agents. Thirdly, conducting problem too complicated narrow AI, i.e., AI proficient dealing only relatively simple well-delineated tasks. If case, should expect fully-fledged until so-called "general" or "human-like" developed. While believe all these ultimately overcome, think being mindful them crucial ensure well-balanced steady progress path psychotherapy.

Language: Английский

---

Evaluating Conversational Agents for Mental Health: Scoping Review of Outcomes and Outcome Measurement Instruments

Journal of Medical Internet Research, Journal Year: 2023, Volume and Issue: 25, P. e44548 - e44548

Published: March 31, 2023

Rapid proliferation of mental health interventions delivered through conversational agents (CAs) calls for high-quality evidence to support their implementation and adoption. Selecting appropriate outcomes, instruments measuring assessment methods are crucial ensuring that evaluated effectively with a high level quality.We aimed identify the types outcome measurement instruments, used assess clinical, user experience, technical outcomes in studies effectiveness CA health.We undertook scoping review relevant literature health. We performed comprehensive search electronic databases, including PubMed, Cochrane Central Register Controlled Trials, Embase (Ovid), PsychINFO, Web Science, as well Google Scholar Google. included experimental evaluating interventions. The screening data extraction were independently by 2 authors parallel. Descriptive thematic analyses findings performed.We 32 targeted promotion well-being (17/32, 53%) treatment monitoring symptoms (21/32, 66%). reported 203 measure clinical (123/203, 60.6%), experience (75/203, 36.9%), (2/203, 1.0%), other (3/203, 1.5%). Most only 1 study (150/203, 73.9%) self-reported questionnaires (170/203, 83.7%), most electronically via survey platforms (61/203, 30.0%). No validity was cited more than half (107/203, 52.7%), which largely created or adapted they (95/107, 88.8%).The diversity choice employed on CAs point need an established minimum core set greater use validated instruments. Future should also capitalize affordances made available smartphones streamline evaluation reduce participants' input burden inherent self-reporting.

Language: Английский

---

Ethical Considerations in Human-Centered AI: Advancing Oncology Chatbots through Large Language Models (Preprint)

JMIR Bioinformatics and Biotechnology, Journal Year: 2024, Volume and Issue: 5, P. e64406 - e64406

Published: Sept. 25, 2024

The integration of chatbots in oncology underscores the pressing need for human-centered artificial intelligence (AI) that addresses patient and family concerns with empathy precision. Human-centered AI emphasizes ethical principles, empathy, user-centric approaches, ensuring technology aligns human values needs. This review critically examines implications using large language models (LLMs) like GPT-3 GPT-4 (OpenAI) chatbots. It how these replicate human-like patterns, impacting design systems. paper identifies key strategies ethically developing chatbots, focusing on potential biases arising from extensive datasets neural networks. Specific datasets, such as those sourced predominantly Western medical literature interactions, may introduce by overrepresenting certain demographic groups. Moreover, training methodologies LLMs, including fine-tuning processes, can exacerbate biases, leading to outputs disproportionately favor affluent or populations while neglecting marginalized communities. By providing examples biased highlights challenges LLMs present mitigation strategies. study integrating human-centric into mitigate ultimately advocating development are aligned principles capable serving diverse equitably.

Language: Английский

---

The Role of Humanization and Robustness of Large Language Models in Conversational Artificial Intelligence for Individuals With Depression: A Critical Analysis

JMIR Mental Health, Journal Year: 2024, Volume and Issue: 11, P. e56569 - e56569

Published: April 27, 2024

Abstract Large language model (LLM)–powered services are gaining popularity in various applications due to their exceptional performance many tasks, such as sentiment analysis and answering questions. Recently, research has been exploring potential use digital health contexts, particularly the mental domain. However, implementing LLM-enhanced conversational artificial intelligence (CAI) presents significant ethical, technical, clinical challenges. In this viewpoint paper, we discuss 2 challenges that affect of CAI for individuals with issues, focusing on case patients depression: tendency humanize lack contextualized robustness. Our approach is interdisciplinary, relying considerations from philosophy, psychology, computer science. We argue humanization hinges reflection what it means simulate “human-like” features LLMs role these systems should play interactions humans. Further, ensuring contextualization robustness requires considering specificities production depression, well its evolution over time. Finally, provide a series recommendations foster responsible design deployment therapeutic support depression.

Language: Английский

---

Countering Misinformation in Private Messaging Groups: Insights From a Fact-checking Chatbot

Proceedings of the ACM on Human-Computer Interaction, Journal Year: 2025, Volume and Issue: 9(1), P. 1 - 30

Published: Jan. 10, 2025

Misinformation on private messaging platforms like WhatsApp and LINE is a global concern. However, research has primarily focused combating misinformation public social media. in difficult to challenge due norms, interpersonal relationships, technological affordances. This study investigates Auntie Meiyu, fact-checking chatbot integrated into LINE, popular service Taiwan. We interviewed 27 users who adopted Meiyu their groups understand motivations perceptions of the chatbot, assess its influence interactions. Participants indicated that they protect close family members from misleading news. Nevertheless, experienced mixed feelings chatbot's robotic style errors detecting misinformation. conclude conversational agents present promising approach for tackling misinformation, particularly when participants disagree, offer design recommendations leveraging AI-enabled countering

Language: Английский

---

Classification of Properties in Human-like Dialogue Systems Using Generative AI to Adapt to Individual Preferences

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(7), P. 3466 - 3466

Published: March 21, 2025

As the linguistic capabilities of AI-based dialogue systems improve, their human-likeness is increasing, and behavior no longer receives a universal evaluation. To better adapt to users, consideration individual preferences required. In this study, relationships between properties human-like system evaluations were investigated using hierarchical cluster analysis for subjects. The driven by generative AI communicated with subjects in natural language via voice-based communication featured facial expression function. Subjective dialogues conducted through questionnaire. Based on results, classified into two types: generally individually relational positive evaluation dialogue. former included inspiration, sense security, collaboration, while latter distance, personality, seriousness. Equipping expected improve most users. should be adjusted individuals since they are evaluated based preferences. A design approach accordance individuality could useful making more comfortable

Language: Английский

---

Attrition in Conversational Agent–Delivered Mental Health Interventions: Systematic Review and Meta-Analysis

Journal of Medical Internet Research, Journal Year: 2023, Volume and Issue: 26, P. e48168 - e48168

Published: Dec. 4, 2023

Background Conversational agents (CAs) or chatbots are computer programs that mimic human conversation. They have the potential to improve access mental health interventions through automated, scalable, and personalized delivery of psychotherapeutic content. However, digital interventions, including those delivered by CAs, often high attrition rates. Identifying factors associated with is critical improving future clinical trials. Objective This review aims estimate overall differential rates in CA-delivered (CA interventions), evaluate impact study design intervention-related aspects on attrition, describe features aimed at reducing mitigating attrition. Methods We searched PubMed, Embase (Ovid), PsycINFO Cochrane Central Register Controlled Trials, Web Science, conducted a gray literature search Google Scholar June 2022. included randomized controlled trials compared CA against control groups excluded studies lasted for 1 session only used Wizard Oz interventions. also assessed risk bias using Risk Bias Tool 2.0. Random-effects proportional meta-analysis was applied calculate pooled dropout intervention groups. compare rate narrative summarize findings. Results The systematic retrieved 4566 records from peer-reviewed databases citation searches, which 41 (0.90%) met inclusion criteria. meta-analytic group 21.84% (95% CI 16.74%-27.36%; I2=94%). Short-term ≤8 weeks showed lower (18.05%, 95% 9.91%- 27.76%; I2=94.6%) than long-term >8 (26.59%, 20.09%-33.63%; I2=93.89%). Intervention participants were more likely attrit short-term (log odds ratio 1.22, 0.99-1.50; I2=21.89%) 1.33, 1.08-1.65; I2=49.43%). Intervention-related characteristics higher include stand-alone without support, not having symptom tracker feature, no visual representation CA, comparing waitlist controls. No participant-level factor reliably predicted Conclusions Our results indicated approximately one-fifth will drop out studies. High heterogeneities made it difficult generalize suggested should adopt blended use tracking, active controls rather controls, reduce rate. Trial Registration PROSPERO International Prospective Systematic Reviews CRD42022341415; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022341415

Language: Английский

---