Digital Risks and AI-Enabled Responses in Adolescent Mental Health: A Bibliometric and Science-Mapping Analysis
- 1BARTIN ÜNİVERSİTESİ, SAĞLIK BİLİMLERİ FAKÜLTESİ, HEMŞİRELİK BÖLÜMÜ
Abstract
Objective: This study mapped digital risks and AI-enabled responses in adolescent mental health, quantified publication balance, and tested explicit domain integration within single records. Method: Scopus metadata (frozen August 9, 2026; exported August 15, 2026) covering articles, reviews, and conference papers were analyzed via bibliometric mapping. Temporal trends evaluated 2015–2025 (complete years), reporting 2026 provisionally. Records were classified as Risk-only, Response-only, or Both. Robustness checks included audits, manual validation of Both records, sensitivity analyses, VOSviewer mapping, thematic evolution, burst detection, and network validation. Results: Of 4,804 retrieved records, 1,623 formed the primary corpus (1,335 in 2015–2025: 833 Risk-only, 495 Response-only, 7 Both). The response-to-risk ratio rose from 0.04 (2015) to 1.19 (2025), when response output first surpassed risk. Manual review confirmed only 5 direct bridges in 2015–2025 (0.37%) and 9 overall (0.55%); integration remained <1% across sensitivity analyses. Keyword mapping revealed distinct cyberbullying/victimization, AI/digital-mental-health, and machine-learning/suicide-prediction clusters, with recent growth in AI, digital phenotyping, and LLMs. Conclusion: Although AI response research has expanded rapidly, explicit co-framing of digital risk and AI response within single publications remains rare and independent of 2026 data.
Related
- AI-enabled and comparator digital interventions for adolescent mental health (2020–2025): early detection, human–AI collaboration, and ethical governance—a PRISMA-ScR scoping review
- Post-Traumatic Stress Disorder After Road Traffic Crashes: A Bibliometric and Science-Mapping Analysis
- Global Research on Grief and Contributions from Türkiye: A Comparative Bibliometric Analysis of the Web of Science and ULAKBIM TR Index
- Knowledge structure and emerging trends of CAR-T therapy in breast cancer : A bibliometric analysis of the Web of Science Core Collection (2012–2025)
- Emergency medicine literature in the Web of Science category “emergency medicine”: A bibliometric analysis (1980–2024)
BibTeX
@article{cicek2026digital,
title = {{Digital Risks and AI-Enabled Responses in Adolescent Mental Health: A Bibliometric and Science-Mapping Analysis}},
author = {Ecem Çiçek},
journal = {Psikiyatride Güncel Yaklaşımlar},
year = {2026},
month = oct,
doi = {10.18863/pgy.2019434},
url = {https://doi.org/10.18863/pgy.2019434},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Digital Risks and AI-Enabled Responses in Adolescent Mental Health: A Bibliometric and Science-Mapping Analysis
AU - Ecem Çiçek
PY - 2026
DA - 2026/10/25
JO - Psikiyatride Güncel Yaklaşımlar
DO - 10.18863/pgy.2019434
UR - https://doi.org/10.18863/pgy.2019434
AB - Objective: This study mapped digital risks and AI-enabled responses in adolescent mental health, quantified publication balance, and tested explicit domain integration within single records. Method: Scopus metadata (frozen August 9, 2026; exported August 15, 2026) covering articles, reviews, and conference papers were analyzed via bibliometric mapping. Temporal trends evaluated 2015–2025 (complete years), reporting 2026 provisionally. Records were classified as Risk-only, Response-only, or Both. Robustness checks included audits, manual validation of Both records, sensitivity analyses, VOSviewer mapping, thematic evolution, burst detection, and network validation. Results: Of 4,804 retrieved records, 1,623 formed the primary corpus (1,335 in 2015–2025: 833 Risk-only, 495 Response-only, 7 Both). The response-to-risk ratio rose from 0.04 (2015) to 1.19 (2025), when response output first surpassed risk. Manual review confirmed only 5 direct bridges in 2015–2025 (0.37%) and 9 overall (0.55%); integration remained <1% across sensitivity analyses. Keyword mapping revealed distinct cyberbullying/victimization, AI/digital-mental-health, and machine-learning/suicide-prediction clusters, with recent growth in AI, digital phenotyping, and LLMs. Conclusion: Although AI response research has expanded rapidly, explicit co-framing of digital risk and AI response within single publications remains rare and independent of 2026 data.
SN - 1309-0674
SN - 1309-0658
ER -