Session 1

AI and Innovation in Social Work Practice

P1  Using ChatGPT to Guide Social Work-Centered Prototyping: Translational Research Going Native and Reflections on Interdisciplinary Collaboration

*Kang Sun¹, Xiaohong Chen², Lin Qi³, Longtao He⁴

¹Southern Illinois University, USA; ²Department of Social Work and Social Policy, Nanjing University, China; ³Institute of Economics, Shanghai Academy of Social Sciences, China; ⁴Research Institute of Social Development, Southwestern University of Finance and Economics, China

Abstract

Background and Purpose: As artificial intelligence tools such as ChatGPT become increasingly embedded in everyday practice, social work faces a critical question: how can practitioners move beyond being end-users of technology to becoming co-creators of AI-informed interventions? This study explores a translational approach in which a social work researcher “goes native” into the prototyping process, using AI as both a design partner and a reflective tool. The purpose is to reimagine the role of social workers in an AI-powered landscape by integrating practice knowledge, ethical reasoning, and rapid prototyping.

Methods: This project adopts a practice-based, iterative design framework combining autoethnographic reflection with applied prototyping. Two parallel development tracks were pursued: (1) a chatbot-based training system for future therapists, designed to simulate client interaction and support skill development through structured feedback; and (2) a smart home prototype addressing night wandering among older adults with cognitive impairment, using low-cost sensors devices. Across both tracks, ChatGPT was used to generate design logic, refine interaction scripts, troubleshoot technical challenges, and support interdisciplinary translation between social work concepts and engineering implementation. Field notes, design iterations, and reflective memos were analyzed to identify patterns in how AI mediated the research and development process.

Results: Findings suggest that ChatGPT functioned as a “boundary object” facilitating communication across domains while accelerating the prototyping cycle. Three themes emerged. First, translation of practice knowledge into technical design: social work concepts such as empathy, safety, and relational presence were operationalized into chatbot dialogue structures and sensor-triggered responses. Second, AI as co-designer and constraint: while ChatGPT enabled rapid ideation and reduced technical barriers, it also introduced biases toward generic solutions, requiring active critical engagement by the researcher. Third, reconfiguration of researcher identity: engaging directly in prototyping blurred traditional boundaries between social worker, researcher, and developer, positioning the social worker as an embedded designer within interdisciplinary teams.

Conclusions and Implications: This study argues that the future role of social workers in AI-integrated contexts lies not only in ethical oversight or service delivery, but indesign participation and technological authorship. “Going native” into prototyping allows social workers to shape technologies that are grounded in lived experience and relational practice. However, this shift also demands new competencies, including technical literacy, critical AI engagement, and collaborative fluency across disciplines. The findings highlight the potential for AI tools like ChatGPT to democratize innovation while underscoring the need for reflexive, values-driven use. Social work education and research must therefore expand to include prototyping as a legitimate form of inquiry and intervention development in the AI era.

P2  Developing an AI-Based Prediction and Intervention Model for Self-Harm Risk Among Elementary School Students in Northwest China

*Hongyan Yin¹, Rui Cao¹, Yahui Li¹, Xuemei Wang¹

¹Lanzhou Universiy, Lanzhou, China

Abstract

Background and Objectives: In the Asia-Pacific region, self-harm behaviors are becoming increasingly common among younger children. While existing research has primarily focused on middle and high school students, there is a lack of systematic early warning and intervention models tailored to elementary school students. This study aims to integrate ecosystem theory with artificial intelligence technology to develop a self-harm risk prediction and intervention model suitable for elementary school settings.

Methods: This study employed qualitative methods, conducting in-depth interviews and observations with 30 elementary school students at risk of self-harm and their significant others in Northwest China to analyze the risk factors, developmental pathways, and systemic interaction mechanisms of self-harm behavior. Based on these findings, a six-dimensional framework encompassing “individual, peers, family, school, community, and hospital” was established. Drawing on crisis intervention models, an AI intervention model was designed that breaks down the self-harm process into four stages: latent risk, overt incubation, crisis onset, and recovery. Corresponding data monitoring, risk assessment, and tiered response protocols were developed for each stage.

Results: The AI intervention model developed in this study systematically bridges universal prevention and targeted intervention: at the school-wide level, it implements primary prevention through life education group activities; at the individual level, it enables tiered early warning based on behavioral and linguistic data analysis, automatically triggering systematic response protocols.

Conclusions and Implications: This study established an AI-based self-harm risk prediction and intervention framework, proposing a new paradigm of “universal prevention, precise identification, and coordinated intervention.” Its implementation requires the establishment of cross-system response mechanisms and strict protection of ethical and privacy standards. Future pilot studies should validate the model’s effectiveness, acceptability, and safety.

P3  Research on the Impact of Family Environment on School Bullying and AI-Assisted Social Work Intervention Strategies

Qing Yu¹, *Dr. Ruoshan Xiong¹

¹Department of Social Work, College of Humanities and Social Sciences, Huazhong Agricultural University, Wuhan, China

Abstract

Background and Objectives: School bullying has become a global public issue, posing serious threats to the physical and mental health of adolescents and attracting widespread social concern. Employing a questionnaire survey, this study systematically explores how the family environment influences school b ullying perpetration and victimization, as well as innovative pathways for social work intervention strategies assisted by artificial intelligence.

Methods: Using survey data collected from 2109 middle school students in Chongqing, Zhejiang and Hubei, this study explores the influences of parental parenting behaviors, parental conflict, parental companionship, and childhood abuse on school bullying behaviors under the guidance of family systems theory and parenting style theory.

Results: The findings revealed that authoritarian parenting style was positively correlated with adolescent bullying perpetration and bullying victimization, while authoritative parenting style was negatively related to both; Increased frequency and intensity of parental conflict significantly elevated the risk of adolescent bullying victimization; The absence of parental companionship reduced adolescents' psychological resilience, which in turn increased their likelihood of bullying victimization. Childhood abuse was closely associated with adolescents' school bullying perpetration.

Conclusions and Implications: Based on these findings, this study constructs a layered intervention strategy centered on “prevention-intervention-follow-up support” and innovatively integrates AI technology into each stage. In the prevention stage, AI empowers social workers to achieve multi-source data fusion for risk identification and proactive early warning, enabling limited social work  resources to be allocated in a targeted manner to high-risk families. In the intervention stage, AI provides digital tools such as virtual role-reversal, emotion regulation assistants, and social skills training simulations, which expands social workers’ means for behavior correction and psychological rehabilitation, and enables real-time, personalized case management. In the follow-up stage, AI facilitates data sharing and task tracking in family-school-community collaboration, consolidating the outcomes of family function restoration and establishing a long-term protective network. This study not only enriches the theoretical framework of school bullying research but also provides new perspectives and feasible pathways for the deep integration and practical application of artificial intelligence insocial work practice.

P4  Digitalised Campus Life and Mental Health Support: A Photovoice Study of Students with Lived Experience of Mental Illness

*Peiyun Yang¹

¹Social Work, Shanghai University, Shanghai, China

 Abstract

Background and Purpose: Despite the strengthening of mental health education, screening, and intervention mechanisms in Chinese higher education, depression and anxiety among university students remain a significant concern, while the everyday experiences and support needs of students with lived experience of mental illness are still insufficiently understood. At the same time, digitalised campus environments are reshaping students’ daily lives, communication practices, and recovery processes. This study aims to present the campus experiences of students with lived experience of mental illness and to examine how digital life is embedded in their mental health situations.

Methods: This study was conducted at a university in eastern China and recruited nine students with lived experience of mental illness, all of whom had received a diagnosis. A participatory Photovoice approach was adopted. Participants were invited to take photographs and write reflective journals around four themes: experiences of emotional distress, everyday campus life, the inner world of the self, and seeking connection. After the photography process, participants analysed their photographs using the SHOWED method followed by one-to-one in-depth interviews and focus groups. Data were analysed through thematic analysis, with coding discussed collaboratively by the research team.

Results: Preliminary findings show that digital media were widely present in participants’ representations of campus life and were deeply embedded in their recovery processes. On the one hand, digital media provided some students with lived experience of mental illness with a relatively low-risk channel for self-expression and emotional support. On the other hand, some students with lived experience of mental illness expressed resistance to digital life. The findings suggest that digitalised ways of living may intensify recovery-related difficulties for students with lived experience of mental illness and also reveal mismatches in existing campus support systems.

Conclusion and Implications: The experiences of students with lived experience of mental illness provide an important entry point for understanding and rethinking campus mental health support. Paying greater attention to the fit between the mental health needs of different student groups and digital environments, while making the obscured everyday experiences of students with lived experience of mental illness visible, may provide an empirical basis for public advocacy, the improvement of campus support systems, and the development of peer support networks.

P5  Integrative Clinical Hypnosis and Aromatherapy Group Intervention for Adolescent Depression in Psychiatry: A Pilot Study

* Wei Ren¹, Pei Sun², *Pei Zhang², Minjie An³

¹China Youth University of Political Studies, China; ²City University of Macau, China; ³Beijing Daxing Xin Kang Hospital, China

 Abstract

Background and Purpose: The prevalence of adolescent depression continues to rise, while current outpatient models remain predominantly pharmacotherapy and individual psychotherapy. Clinical hypnosis can reshape cognition and regulate emotion through suggestion and imagery; aromatherapy acts directly on the limbic system via olfactory pathways. Integrating both within a group setting can simulate the core functions of a healthy family (safety, somatic comfort, and positive connection), thereby reconstructing psychological safety. This study hypothesized that simultaneously engaging five channels (olfactory, tactile, auditory, visual, and interpersonal) would significantly reduce depressive symptoms, and aimed to evaluate the effectiveness of this integrated group intervention.

Methods: A single-group pre-post design was adopted. Six adolescents aged 12–18 years were recruited from a psychiatric outpatient clinic to receive a single 2.5-hour integrated group session led by a practitioner with specialized training in clinical hypnosis and aromatherapy. The intervention integrated hypnotic metaphorical stories, aromatherapy inhalation, peer hand massage, and structured interpersonal interaction across four phases: induction, deepening, exploration, and integration. Depressive symptoms were assessed using the Beck Depression Inventory-II (BDI-II) before and after the intervention, with data analyzed using paired-samples t-test.

Results: Post-intervention BDI-II total scores demonstrated a statistically significant decrease, with pre-test mean of 36.50 (SD = 10.60) and post-test mean of 22.33 (SD = 15.62), t(5) = 2.85, p = 0.036, Hedges` g = 1.07, indicating a large effect size and a 38.81% reduction.

Conclusions and Implications: The integrated intervention significantly alleviated depressive symptoms in adolescents and may serve as an effective adjunctive therapy within hospital systems, providing an evidence base for an integrated service model centered on "reconstructing a psychologically safe home." Although not directly tested against digital interventions, its core advantages lie in the irreplaceability of multi-sensory integration and authentic therapeutic relationships: in-person aromatherapy hand massage evokes early secure attachment through real interpersonal contact, guiding adolescents back to authentic connection; the hypnotherapist can also adjust guidance in real time based on in-session dynamics, allowing for flexible focus and attunement. At the practice and policy level, the "psychologically safe home" framework could inform parenting education programs to address the root causes by strengthening caregiving capacities from the source. Future research should validate the effectiveness through large sample randomized controlled trials, incorporate biological markers such as cortisol and heart rate variability, and develop longer-term session series to examine sustained effects.

P6  Effects of Mindful Painting Group Intervention on Anxiety and Peer Relationship Among Adolescent Girls Living in Single-Parent Families in Macao

Chao U¹, *Donghang Zhang², Jierong Hu², Waisan Wong¹

¹ Institute of Analytical Psychology, Faculty of Health and Wellness, City University of Macau, China; ²Department of Innovative Social Work, Faculty of Health and Wellness, City University of Macau, China

Abstract

Adolescent girls from single-parent families often face psychosocial challenges. The mindful painting group intervention has gained increasing attention in recent years but lacks adequate research, especially in China. This study aimed to evaluate the effect of this approach on anxiety, active procrastination and peer relationships among Chinese adolescent girls. Using mixed methods and a quasi-experimental design, this study involved 20 teenage girls from single-parent families, with 10 girls each assigned to either the treatment or control groups. Participants in the treatment group received 8 weeks of mindful painting intervention, while participants in the control group were treated as usual. The quantitative findings revealed a significant decrease in anxiety and a significant increase in active procrastination and peer relationship quality in the treatment group before and after the intervention. No significant change was found in the control group. Furthermore, qualitative findings revealed participants' active expression of academic stress and anxiety, promotion of peer relationships and support, enhancement of self-awareness and acceptance and reduction of anxiety. The findings provide empirical evidence for social workers to adopt the mindful painting intervention for addressing psychosocial challenges and a group collaborative approach to facilitate the exchange of participants' experiences and enhance emotional responsiveness among them.

P7  A Qualitative Study in Accessing the Needs, Acceptance, and Preferences of AI-Based Conversational Agents among Older Chinese Immigrants

*Xiayu Summer Chen¹, Flavia Andrade², Qingwen Xu³

¹School of Social Work, University of Central Florida, Flordia, USA; ²School of Social Work, University of Illinois Urbana Champaign, Illinois, USA; ³School of Social Work, New York University, New York, USA; Graduate Programs, New York University Shanghai, Shanghai, China

 Abstract

Background and Purpose: Older Chinese immigrants in the U.S. face unique challenges to aging in place, including language barriers, cultural differences, and social isolation. While AI-based conversational agents (CAs) offer potential for health support and information seeking, research on how to tailor these tools for culturally and linguistically diverse populations is scarce. This study explores the specific needs, acceptance, and design preferences of older Chinese immigrants regarding AI-based CAs to facilitate healthy aging in place.

Methods: This exploratory qualitative study utilized semi-structured, in-depth interviews with 13 older Chinese immigrants (aged 60 and above). Interviews were conducted in Mandarin, Cantonese, or English to ensure linguistic comfort. Participants were shown demonstrations of AI applications to ground their perceptions. Data were analyzed using thematic analysis to identify key recurring patterns in participants' needs, perceived barriers, and design preferences.

Results: The findings indicated that participants have significant needs for health management support, such as medication reminders, as well as for social connectedness and assistance with reliable digital information seeking. While attitudes toward AI ranged from cautious curiosity to enthusiastic engagement, trust was found to be highly contextual; participants felt comfortable using CAs for low-risk informational tasks but expressed significant caution regarding high-stakes medical advice. Foundational barriers to adoption included low technical self-efficacy, limited digital skills, and persistent concerns regarding privacy, data security, and the potential for scams. Regarding design, participants emphasized a strong preference for voice-based interaction and a warm, empathetic conversational tone rather than didactic or text-only interfaces. Furthermore, the analysis highlighted the necessity of "cultural fluency," where agents must demonstrate an understanding of Chinese values, idioms, and dialects to establish a sense of respect and social inclusion.

Conclusions and Implications: The study underscores the potential of AI-based CAs to act as inclusive tools for aging in place, provided they are developed with cultural sensitivity and linguistic accuracy. These results suggest that digital interventions must move beyond functional innovation to incorporate relational and cultural depth. Social work practice and policy should prioritize community-led digital literacy training and the development of multilingual, equitable AI technologies that address the specific vulnerabilities of immigrant older adults to ensure technology reduces rather than widens existing health disparities.

P8  Reconstructing Human Subjectivity in AI-Assisted Home Visit Practice: A Reflective Learning Model Intervention Study

*Fan Xinyue¹, Shen Qing², Li Yingqi³, Zhang Fangfang⁴

¹Luoke Social Service Centre, Beilun, Ningbo, China; ²Luoke Social Service Centre, Beilun, Ningbo, China; ³Society Work Department, Beilun, Ningbo, China; ⁴Beilun Volunteer Association, Beilun, Ningbo, China

Abstract

Background and Purpose: As artificial intelligence (AI) systems become increasingly embedded in frontline social work practice, questions have arisen regarding how professional subjectivity is sustained, reshaped, or constrained within AI-mediated workflows. This study examines how structured reflective supervision influences human subjectivity among social workers who utilize AI-generated home visit outlines. Drawing on the Reflective Learning Model, the study conceptualizes subjectivity as a developmental and reconstructible professional capacity grounded in autonomy, ownership, competence, and reflective awareness.

Methods: A quasi-experimental longitudinal design was implemented within a social service organization where AI-generated home visit outlines are routinely employed. Participant recruitment has been completed, and baseline assessments are ongoing. Baseline measures include self-reported human subjectivity in AI usage, perceived professional agency, autonomy orientation, and attitudes toward AI integration.

Results: Social workers were divided into two groups based on baseline subjectivity orientation: those experiencing tension when using AI and those reporting neutrality. Both groups participated in an intervention consisting of structured practice logs aligned with the Reflective Learning Model. Participants document AI-related events (Event), reflect on emotional and professional impact as well as theoretical and procedural implications (Exploration), articulate planned adjustments in practice (Experimentation), and subsequently evaluate outcomes (Evaluation). Additionally, behavioral data are drawn from system-generated AI usage logs, capturing frequency of AI consultation, patterns of outline modification, and structural divergence between AI-generated and finalized home visit documentation. Pre-and post-intervention measures will assess changes in self-reported subjectivity, perceived autonomy, and behavioral modification intensity. Comparative analyses will examine whether subjectivity is reinforced, diminished, or reconfigured through guided reflection.

Conclusions and Implications: This study advances four key contributions. First, it reframes human subjectivity in AI-assisted professional contexts as a dynamic and reconstructible capacity rather than a static trait. Second, it integrates psychological theory with behavioral trace data, linking subjective experience, reflective practice, and actual AI modification behavior. Third, findings may inform the iterative development of home visit AI systems by identifying interaction patterns that foster autonomy-supportive engagement rather than passive reliance or defensive resistance. Fourth, beyond system design implications, the study offers potential pathways for enhancing professional agency in the AI era and explores the feasibility of forecasting or nowcasting human subjectivity through AI usage.

P9  Developing and evaluating a family–school–community collaborative model to children weight management: An experimental study

*Meng Cao¹, Shitong Shao¹

¹Faculty of Health and Wellness, City University of Macau, China

Abstract

Purpose: Childhood obesity has become a global public health concern. In China, the combined rate of overweight and obesity among children stands at 19.9%, highlighting the urgency of effective prevention. The National Health Commission and relevant departments have launched the "Year of Weight Management" Implementation Plan, targeting a 70% reduction in childhood obesity over the next decade. Experts agree that collaboration among families, schools, and communities is essential to improving child health. The Theory of Overlapping Spheres of Influence (TOSI) provides a theoretical basis and operational framework for such cooperation. Guided by TOSI, this study carries out a comprehensive intervention trial to develop a family–school–community (FSC) collaborative model for obesity prevention and management. The aim is to create a replicable strategy that can be applied in various settings, ultimately enhancing the physical health of obese children.

Methods: A total of 60 obese children (aged 12.4 ± 0.7 years; 32 boys and 28 girls) were recruited and randomly assigned to either the intervention group (IG, n=30) or the control group (CG, n=30). The intervention comprised three components: family-based measures, including health education, physical activity assignments, and parent–child exercises; school-based measures,consisting of a comprehensive curriculum, individualized exercise prescriptions, and dietary guidance; and community-based measures, such as public awareness lectures, teacher consultations, and community sports programs. Body composition (such as BMI, body fat percentage (BF%), fat mass (FM), visceral adipose tissue area (VAT)), cardiorespiratory fitness (maximal oxygen uptake (VO2max)), and glycolipid profiles (fasting glucose (FBG), total cholesterol (TG), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C)) indicators were measured for all participants both before and after the intervention. Data were analyzed using SPSS software to assess within-group changes and between-group differences in these outcomes.

Results: After six months of comprehensive intervention, compared with before intervention, the body weight (-3.3 kg), BMI (-1.7 kg/m2), BF% (-3.4%), FM (-3.3 kg), VAT (-20.8 cm2), and FBG (-0.27 mmol/L) of children with obesity in the IG group were significantly reduced (P<0.05), while the VO2max in both IG and CG groups were significantly improved (+3.1 mL/kg/min and +1.6 mL/kg/min, respectively, P<0.05). The blood lipid indicators of both groups showed significant changes. Compared with the CG group, the IG group showed greater improvements in body composition, cardiorespiratory fitness, and fasting blood glucose in children with obesity (P<0.05).

Conclusions and Implications: This study first confirmed the feasibility of a family–school–community (FSC) collaborative model in the prevention and management of adolescent obesity. The intervention results demonstrated that measures implemented separately in the school, family, and community setting swere all effective. The FSC model significantly improved cardiorespiratory fitness in obese children and also led to favorable changes in blood glucose levels. However, no significant improvement was observed in lipid profiles. This may be primarily attributed to the absence of strict dietary restrictions during the intervention, as well as the inclusion of school summer and winter vacations, which likely caused periodic fluctuations in participants’ dietary and physical activity patterns.

P10  Analysis of the User Experience and Intervention Mechanisms of AI Smoking Cessation Tools: A Qualitative Study

*Yufei Xia¹, Xue Weng²

¹School of Sociology, Beijing Normal University, Beijing, China; ²Institute of Advanced Studies in Humanities and Social Sciences, Beijing Normal University, Zhuhai, China

 Abstract

Objective: Smoking rates are relatively high among university students. Despite the availability of various smoking cessation services (such as smoking cessation clinics, medications, and helplines), the acceptance and participation rates of university students in these services remain low. Increasing the utilization of these services has become a critical issue in public health and campus wellness. This study explores the potential of AI-based smoking cessation tools, particularly within university populations.

Methods: This study used purposive sampling to select 14 participants from the smoking cessation activity in the smoke-free campus initiative at the Zhuhai campus of Beijing Normal University. These participants had used an AI tool based on positive psychology for smoking cessation. Semi-structured interviews focused on the feasibility and effectiveness of the tool, with data analyzed through thematic analysis.

Results: The AI tool facilitated three key mechanisms in smoking cessation. First, it provided timely, evidence-based cessation advice, enhancing participants' knowledge. Second, it utilized positive psychology techniques (e.g., motivational language and achievement recognition) to increase self-efficacy and motivation. Third, the tool applied emotional regulation strategies, improving psychological resilience and helping participants manage stress during the cessation process.

Conclusions and Implications: AI-based smoking cessation tools effectively boost motivation and success rates among university students by offering scientific advice, positive reinforcement, and emotional support. Their convenience and 24/7 availability make them a valuable resource for continuous and flexible support in smoking cessation. Further development and promotion of such tools on campuses are recommended to enhance adolescent smoking cessation efforts.

P11 A Randomized Controlled Trial to Enhance Financial Literacy among Yi Adolescents in Relocation Communities

*Siyu Long¹

¹Central University of Finance and Economics, Beijing, China

Abstract

Background and Purpose: Relocation poverty alleviation policies in China have shifted many Yi minority adolescents from rural environments to urban communities, exposing them to increasingly complex financial decision-making contexts. While existing research has primarily focused on educational adaptation and behavioral issues among relocated youth, limited attention has been paid to their financial capability development, particularly within culturally responsive intervention frameworks. Financial literacy—conceptualized as an integrated construct encompassing financial knowledge, behaviors, and attitudes—represents a critical form of developmental capital for adolescents navigating financialized societies. This study constructs a culturally responsive, experiential learning–based group work model and examines its effectiveness in enhancing financial literacy among Yi adolescents in relocation communities through a randomized controlled trial (RCT).

Methods: An RCT was conducted in 2024 in a Yi relocation community in Liangshan, Sichuan Province. Twenty-two adolescents (aged 12–17) who met inclusion criteria were randomly assigned to an intervention group (n=12) or a control group (n=12). The intervention group participated in a nine-session structured group program integrating financial literacy theory with culturally responsive and experiential learning principles. The program targeted three domains: financial knowledge acquisition, behavioral skill training, and attitude formation. The control group received no thematic intervention. Financial literacy was measured using a culturally adapted version of the Youth Financial Literacy Scale. Given the small sample size and non-normal distribution, Mann–Whitney U tests and Wilcoxon signed-rank tests were conducted to assess between-group and within-group differences, and effect sizes were calculated.

Results: No significant differences were observed between groups at baseline (p > .05). Post-intervention analyses indicated that the intervention group scored significantly higher than the control group (p < .001). Within-group comparisons showed a significant improvement in financial literacy among participants receiving the intervention (p < .01), with a large effect size (Cohen’s d = 1.70), while no significant changes were found in the control group.

Conclusions and Implications: Findings demonstrate that a culturally responsive, experiential group intervention can effectively enhance financial literacy among minority adolescents in relocation communities. This study contributes empirical evidence to culturally grounded social work interventions in ethnic minority contexts and provides a replicable practice model for promoting youth financial capability. Future research should expand sample sizes an incorporate longitudinal designs to examine the sustainability and underlying mechanisms of intervention effects.

P12  AI and Disabilities: Pros and Cons

*Violet E. Horvath¹

¹Pacific Disabilities Center, John A. Burns School of Medicine, University of Hawaii at Manoa, Honolulu, Hawaii, USA

Abstract

Background and Purpose: While the use of AI is surging, much of it began as assistive technology for persons with disabilities. Today, many assistive technologies designed for those with disabilities are now widely used by and benefit everyone. This presentation covers the pros and cons of AI and disabilities, and their implications.

Methods: Many sources were consulted, including peer-reviewed journal articles, newspaper articles, government and organization websites, and social media posts.  The pertinent information was synthesized and includes, among other things, a definition of AI, types and concepts of AI, that AI is assistive technology and examples, and the drawbacks.

Results: AI-powered assistive technology has been around since 1976, with the first “reading machine.” Today, there are many text-to-speech apps and software available to everyone. AI assistive technologies and devices include real-time captioning, navigation systems, early detection of health issues, chatbots that deliver reminders, and the translation of documents into different languages. Problems arise when AI is used to replace humans – when it replaces social workers, mental health professionals, physicians, and others. AI systems are designed around the needs of the majority, and include little to no information on those with disabilities and those from other marginalized communities. Automation is meant to remove human bias and discrimination, but instead, it promotes it. Many are turning to unregulated AI therapy bots because of a lack of human therapists, sometimes resulting in serious complications, including suicide. These are but some of the issues surrounding AI.

Conclusions and Implications: AI assistive technology is a powerful tool for those with and without disabilities. In general, AI thrives in its role of assisting humans. Drawbacks occur when AI is used to replace humans, as there is little to no control or guardrails, creating new and serious problems we are ill-equipped to deal with.

P13  The Application of Artificial Intelligence in Social Work: A Scoping Review

*Jie Song¹, Bo Fu¹, Xinghua Ren¹, Yunan Wu¹, Zheng Xu¹

¹School of Philosophy and Sociology, Lanzhou University, Lanzhou, China

Abstract

Background & Purpose: As artificial intelligence (AI) applications such as machine learning, large language models, and conversational agents become more prevalent in healthcare, education, and public services, their potential integration into social work practice warrants careful examination. To systematically review the current state of research on artificial intelligence in social work, identify the main application scenarios, effectiveness, and limitations of AI technologies, and provide references for future research and practice.

Methods: A scoping review was conducted by systematically searching databases including Web of Science, Embase, Cochrane Library, APA PsycInfo, and Ovid MEDLINE(R) ALL. Literature was screened according to inclusion and exclusion criteria, and thematic analysis was performed from the perspectives of AI technology types, application scenarios, and effectiveness evaluation.

Results: A total of 30 studies were included. The findings indicate that AI technologies in social work are primarily applied in the following areas: education and training, assessment and diagnosis, intervention and mental health support, decision support, data and text analysis, and ethical and practice reflection. The main forms of AI include generative AI and large language models, chatbots and virtual assistants, machine learning and predictive models, and human-computer interaction. AI technologies show promise in enhancing information collection, providing basic psychological support, expanding service accessibility, and assisting in risk assessment, screening, and prediction. However, limitations remain in areas such as understanding complex human needs and emotions, cultural adaptability, ethical concerns, and data privacy and security.

Conclusions and Implications: The application of AI in social work is developing rapidly, showing broad potential in education, assessment, intervention, and management. Current research is characterized by three main features: diversified applications, accelerated technological iteration, and reflective development. Future directions should include strengthening interdisciplinary collaboration, promoting deeper integration of AI with professional ethics, emphasizing local adaptation, conducting high-quality empirical studies to verify long-term effectiveness and fairness, and seeking a balance between technology and humanistic care.

P14  Digital Intervention of Social Prescribing in Addressing Elderly Loneliness: A Literature Review

* Kaihe Zhang¹, Ying Wang¹,

¹School of Philosophy and Sociology, Lanzhou University, Lanzhou, China

Abstract

Background & Purpose:  Social work practice has undergone rapid digitalization. However, social work education has not kept pace: most Master of Social Work (MSW) curricula presume face-to-face interaction as the default context, leaving students unprepared for technology-mediated practice. When students enter online practicum placements, they encounter concrete dilemmas: How do I build trust without seeing facial expressions? How do I balance client expectations with my own sustainability? How do I convey genuine understanding across a screen?

Methods: This teaching note presents three integrated modules grounded in analysis of 132 real-world online client-worker interactions and feedback from 30 MSW students across a 6-month practicum. Rather than treating online service delivery as a "simplified version" of face-to-face practice, the framework proposes "relational co-presence"—a new form of empathy grounded in classical Chinese philosophy (Zhuangzi) and emphasizing shared meaning-making rather than physical proximity.

Results: The core framework proposes a three-dimensional trust cycle: System Trust  Relational Space → Interpersonal Trust, operationalized through strategic use of "we-language," emoji calibration, plain language, and transparent communication delays. All three modules are aligned with CSWE 2022 EPAS competencies and require 60-75 minutes of classroom time.

Results demonstrate significant effectiveness: 70% of students mastered "we-language" techniques; 80% accurately calibrated emoji use; digital empathy self-efficacy improved significantly (paired t(27) = 8.34, p < 0.001, d = 1.53); post-practicum supervisor ratings showed 23 percentage-point improvements compared to controls.

Conclusions and Implications: The teaching note provides a theoretically-grounded framework, three immediately implementable classroom modules with concrete examples and assessment rubrics, empirical evidence of effectiveness, and comprehensive appendices. This reimagining of empathy does not abandon social work's relational soul in the digital era—it is a matured form of professional practice.

P15  AI-Enabled Emotional Companion Intervention for Elderly Loneliness: An Empirical Study Based on Nanjing Baishan Social Work Practice

*Zhijing Wu¹

¹School of Philosophy and Social Sciences, Lanzhou University, Lanzhou, China

Abstract

Background and Purpose: Loneliness is a prevalent and persistent issue among older adults, severely affecting their physical and mental health, daily functioning, and overall quality of life while also increasing the risk of chronic diseases and psychological distress. With the rapid acceleration of population aging in China, traditional elderly social work models are increasingly facing prominent challenges, including insufficient professional human resources, limited service coverage in remote or community-based settings, and difficulty meeting the personalized emotional needs of the elderly. To address these gaps, this study explored the effectiveness of AI-enabled emotional companion interventions in alleviating elderly loneliness, taking the practical experience of the Nanjing Baishan Social Work Team as a typical case. It hypothesized that AI digital human companion systems, when closely integrated with professional social work services, could significantly reduce the level of loneliness among older adults.

Methods: A pre-post experimental design was adopted for this study, with 86 elderly residents from 3 pilot nursing homes in Nanjing selected as participants, ranging in age from 65 to 88 years old, including 47 females and 39 males. Data were collected through multiple methods, including the UCLA Loneliness Scale (a widely recognized tool for measuring loneliness), semi-structured interviews to capture subjective experiences, and system usage logs to track interaction patterns. The intervention lasted for 3 months, utilizing Baishan’s self-developed AI digital human system, which is equipped with advanced speech and emotion recognition technologies as well as life story reproduction functions to better resonate with the elderly; descriptive statistics and paired sample t-tests were employed for systematic data analysis.

Results: After the 3-month intervention, the average loneliness score of the participants decreased significantly from 42.3 to 24.8 (t=11.26, p<0.001), representing a 41.4% reduction in overall loneliness levels. Specifically, 82.6% of the participants reported a noticeable reduction in their feelings of loneliness, and 76.7% stated that the AI companion system effectively alleviated their sense of isolation and emptiness in daily life. System usage logs further showed that participants had an average daily interaction time of 48 minutes with the AI system, and professional social workers conducted targeted emotional intervention in 19 cases where the system detected negative emotions such as sadness or anxiety.

Conclusions and Implications: The findings indicated that AI-enabled emotional companion interventions are highly effective in reducing elderly loneliness when integrated with professional social work services. The case of Nanjing Baishan Social Work Team demonstrated that AI can effectively extend the boundaries of elderly social work services and complement human care, especially in scenarios where human resources are scarce. Practically, this study provides a scalable and replicable model for elderly social work practice; policy-wise, it offers empirical support for promoting the application of AI technologies in community elderly care services. Future research should focus on exploring the long-term effects of such interventions and optimizing the cultural adaptability of AI systems to better meet the needs of older adults.

P16  Artificial Intelligence and the Occupational Well-being of University Student Affairs Practitioners: Potential Impacts and Ethical Considerations

*Xinghua Ren¹, Jie Song¹, Jiyixi Hua¹, Tianming Bai¹, Bo Fu¹, Yaling Shang¹, Zheng Xu¹

¹School of Philosophy and Social Sciences, Lanzhou University, Lanzhou, China

Abstract

Background and Purpose: Artificial intelligence is increasingly being integrated into university student affairs, with applications spanning AI-assisted counselling, behavioural analytics, and automated administrative processes. Recent international surveys indicate that a considerable proportion of university faculty and staff experience pressure and uncertainty when engaging with AI tools, often operating in a state of “learning by doing.” However, existing scholarship has predominantly focused on technological effectiveness or student-level outcomes, leaving the occupational well-being of practitioners themselves underexplored. In this context, student affairs professionals—as direct users of AI and as providers of relational services—face emerging challenges that warrant systematic examination. This paper aims to investigate the potential implications of AI applications for the occupational well-being of university student affairs practitioners and to critically examine the key ethical dimensions involved.

Methods: This paper adopts conceptual analysis and a literature review approach. It identifies salient AI application scenarios within university student affairs and, drawing upon the Professional Quality of Life (ProQOL) framework and theories of technology ethics, systematically analyses the potential effects of AI on practitioners’ occupational well-being as well as the associated ethical concerns.

Results: The findings suggest that AI applications in student affairs carry a dual character. On one hand, AI can assume routine administrative tasks, thereby enhancing efficiency and reducing workload burdens. On the other hand, the associated risks are more complex. First, erosion of professional autonomy—an over-reliance on algorithmic decision-making may narrow the scope for professional discretion and weaken practitioners’ sense of professional identity. Second, alienation of emotional labour—when emotionally intensive work becomes standardised or substituted by technology, practitioners may experience role ambiguity and a diminished sense of meaningfulness. Third, blurring of work–life boundaries—the continuous accessibility of intelligent systems exacerbates the dissolution of boundaries between professional and personal life. Fourth, unclear ethical accountability—when technology-assisted decisions lead to adverse outcomes, attribution of responsibility becomes ambiguous, potentially heightening occupational anxiety.

Conclusions and Implications: In advancing the digitalisation of student affairs, universities should incorporate practitioner occupational well-being into their technology assessment frameworks. Recommendations include establishing institutional mechanisms for ethical review and impact assessment; clarifying the boundaries and accountability structures of AI-assisted decision-making while preserving space for professional judgment; and providing adequate training and technical support to enable practitioners to transition from passive use to active mastery. Sustained attention to the human dimension alongside technological empowerment is essential to achieving a balanced outcome between efficiency gains and occupational well-being.