No open-access location is known for this record yet.
No open-access location is known for this record yet.
Adolescent mental health is increasingly affected by complex stressors across academic, social, familial, and identity-related domains. With the rise of generative artificial intelligence (AI), tools such as ChatGPT, Gemini, and DeepSeek offer new pathways for psychological support. This study investigates the impact of AI-based interventions on adolescent mental health across multiple domains, including academic stress, peer relationships, family dynamics, gender identity issues, financial concerns, and physical and psychological well-being. A cross-sectional survey was conducted among 217 Thai high school students aged 15–18. Participants responded to a structured questionnaire assessing their stress levels, emotional needs, and usage of AI platforms for support. Likert-scale metrics were used to evaluate perceived effectiveness and satisfaction across each domain. Statistical analyses, including correlation, t-tests, and ANOVA, were employed to explore relationships between stress factors, demographic variables, and perceived AI impact. Results indicate that while no statistically significant differences were found in AI satisfaction between gender or grade levels, students experiencing high academic or social stress tended to report slightly higher engagement with AI tools. Notably, AI was perceived as most helpful in academic support and emotional ventilation, especially among students facing peer or family-related stress. This research contributes to the understanding of AI’s potential as an accessible mental health intervention for adolescents. It highlights the multidimensional roles generative AI tools can play in supplementing emotional well-being support, while also emphasizing the need for further longitudinal and ethically guided studies.
Loading…
Stored only when both papers are in the corpus — 2 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.