The Impact of ChatGPT-Assisted Emotional Regulation on Chinese Middle School Students’ Social-Emotional Competencies

• In Event: Roundtable Session 4
  In Refereed Roundtable Session: Country cases studies on the challenges and opportunities of education and technology integration

Proposal

Abstract
Based on the Organization for Economic Co-operation and Development (OECD) framework for social and emotional competencies, this study aims to investigate the effects of ChatGPT-assisted emotional regulation interventions on the social-emotional competencies of Chinese middle school students, specifically stress resistance, optimism, and emotional control. Employing a mixed-methods approach, the study uses convenience sampling to select 200 ninth-grade students from five middle schools in Chongqing. The intervention includes a one-hour ChatGPT-assisted emotional regulation activity, encompassing simulated scenarios and feedback exercises. Data collection involves emotional regulation questionnaires, focus group discussions, and semi-structured interviews. This study aims to provide empirical evidence for designing and implementing AI-assisted social-emotional learning (SEL) curricula and offer insights for educators and policymakers to promote SEL among Chinese middle school students.

Objective
Since 2010, China has emphasized middle school students’ social-emotional development, recognizing their dual challenges of academic pressure and adolescent psychological changes (Wang & Tang, 2022; Fang, 2023). Despite SEL’s importance, it’s often overlooked due to exam-oriented education(Shi & Li, 2013; Mao, 2019), requiring innovative SEL teaching models. Meanwhile, technological advancements, especially AI, are transforming education by revolutionizing teaching methods, learning environments, and student outcomes (Russell & Norvig, 2022). Although AI also demonstrates the potential to enhance SEL skills (Surbhi & Kanishk, 2024), it faces challenges like data privacy and ethics (Tene & Polonetsky, 2012), warranting further research on its SEL applications.
This study seeks to understand ChatGPT’s potential and limitations in promoting emotional regulation in Chinese middle school students, providing empirical evidence for AI integration in social-emotional learning. The research questions are: (1)How does ChatGPT-assisted AI emotional regulation intervention enhance students’ stress resistance, optimism, and emotional control abilities? (2) What difficulties and challenges do middle school students encounter when using ChatGPT for emotional regulation training?

Theoretical Framework
This research is grounded in the OECD’s social-emotional skills framework. The framework, derived from the “Big Five Model,” categorizes skills into five dimensions: task performance, emotional regulation, collaboration, open-mindedness, and engagement with others (Huang, 2020). Emotion regulation, focusing on stress resistance, optimism, and emotional control, is crucial for middle school students to deal with academic pressures and adolescent mood fluctuations (Ivcevic & Brackett, 2014). AI technology, especially ChatGPT, has the potential for widespread application in education by helping students practice emotion recognition and regulation through personalized, instant feedback (Demetriou et al., 2022). Therefore, this study chooses emotion regulation as the core intervention target and combines three abilities from the OECD framework for assessment: stress resistance, optimism, and emotional control.

Literature Review
Application of Generative Chatbots in Education
AI technology has made significant educational strides, particularly in personalized learning, intelligent tutoring systems, and adaptive learning environments (Luckin & Holmes, 2016). Among various AI tools, generative AI chatbots have gained attention for their ability to produce coherent, systematic, and informative responses (Zhai, 2023). ChatGPT, for instance, provides students a non-judgmental environment, encouraging emotional expression and reflection (Amershi et al., 2019). Despite its potential, ChatGPT faces challenges in educational implementation. Key concerns include data privacy and security (UNESCO, 2021), algorithmic bias in different cultural contexts (Zhao & Vandenberghe, 2020), and the potential to weaken students’ independent learning and critical thinking abilities (Wogu et al., 2018).
Social-Emotional Development of Chinese Middle School Students
Middle school students encounter substantial social-emotional challenges stemming from developmental inconsistencies (Han et al., 2021), manifesting as social anxiety, low confidence, and public speaking fear, potentially impacting self-esteem and academic achievement (Wang & Tang, 2022). Empirical studies indicate a higher prevalence of mental health issues, particularly depression, among Chinese adolescents compared to European counterparts, especially in grades 9-11 (Zhou et al., 2023). In response, SEL has gained traction in Chinese education. Introduced officially in 2011 through Ministry of Education and UNICEF pilot projects across multiple provinces, SEL implementation in China progressed significantly from 2011 to 2021 (Shi & Li, 2013). Key developments include the formulation of a “three-dimensional six-aspect” conceptual framework and culturally aligned curricula (Wang & Tang, 2021). Despite progress, SEL implementation in China faces persistent challenges. Exam-oriented education systems impede the allocation of resources to SEL programs (Mao, 2019). Moreover, insufficient teacher training hinders effective SEL integration into daily pedagogical practices (Shi & Li, 2013).

Methodology
This study examines 200 ninth-grade students from five Chongqing middle schools, selected via convenience sampling to ensure diverse representation and the need for prior school engagement and consent. A mixed-method approach is employed, combining quantitative and qualitative analysis to assess the impact of a 1-hour ChatGPT-assisted emotion regulation intervention on students’ stress resistance, optimism, and emotional control. The intervention consists of three independent modules, each focusing on one of these dimensions, with simulations of academic stress, interpersonal conflicts, and related exercises. Afterward, students complete a self-assessment questionnaire using a validated 5-point Likert scale (Cronbach’s α > 0.80). The research team will also conduct 45-minute focus group discussions and 30-minute semi-structured interviews (voluntary) to identify common themes and gather detailed feedback on the intervention.

Data Analysis
SPSS 27.0 and R 4.1.0 will be employed for inferential statistics. Paired sample t-tests will compare pre- and post-intervention scores across stress resistance, optimistic attitude, and emotional control. RM-ANOVA will analyze intervention effects over time. Cohen’s d will quantify effect sizes. Pearson correlations will examine interrelationships among the three dimensions. NVivo 12 will facilitate coding and theme extraction from the interview and focus group transcripts, focusing on students’ experiences and suggestions. Content analysis of student-ChatGPT dialogues will assess interaction quality and effectiveness.

Result
This study evaluates the impact of ChatGPT-assisted emotion regulation interventions on Chinese middle school students’ stress resistance, optimism, and emotional control. Findings indicate enhanced stress coping through simulated scenarios and strategy provision, improved optimism via positive feedback and cognitive restructuring exercises, and increased emotional control through dialogues and strategy learning. Qualitative analysis reveals students’ experiences and suggestions, highlighting ChatGPT’s strengths (e.g., immediate feedback, non-judgmental environment) and potential limitations (e.g., insufficient dialogue depth). Results demonstrate ChatGPT’s effectiveness in emotion regulation, prompting further study on AI-assisted strategies for holistic adolescent development.

Author

• Yanling Chen, Michigan State University