EUVIMEDEUVIMED entdecken
Mixed-Methods-StudieOpen Access

Depictions of Depression in Generative AI Video Models: Mixed Methods Study of OpenAI’s Sora 2

Matthew Flathers, Griffin Smith, Julian Herpertz, Zhitong Zhou, John Torous

Journal of Medical Internet Research · 14.09.2026 · 28 · e95682-e95682

DOI: 10.2196/95682 ↗

Kurz gefasst

Mixed-Methods-Studie in der Zeitschrift Journal of Medical Internet Research, online erschienen am 14.09.2026. Verfasst von Matthew Flathers und 4 weiteren Autor:innen. Beteiligte Einrichtung: Division of Digital Psychiatry, Beth Israel Deaconess Medical Center, 20 Overland St, Suite OV-202, Boston, MA, 02215, United States, 1 6176676700. Hauptthemen laut MEDLINE: Depression, Video Recording. Die Arbeit nennt 44 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Generative AI video models are increasingly capable of producing complex depictions of mental health experiences, yet little is known about how these systems represent conditions such as depression. Because AI-generated content may reach people during vulnerable periods, understanding what visual narratives these models produce for sensitive concepts carries clinical relevance.

Objective

This study aimed to characterize how OpenAI’s Sora 2 generative AI video model depicts depression and examine whether depictions differ between the consumer app and developer API access points, which differ in their product layer mediation.

Methods

We generated 100 videos using the single-word prompt “Depression” across 2 access points: the consumer app (n=50, 50%) and developer API (n=50, 50%). Two trained coders independently coded narrative structure, visual environments, objects, figure demographics, and figure states. Interrater reliability was assessed using the Cohen κ, with dimensions showing insufficient agreement excluded from analysis. Computational features (visual aesthetics, audio, semantic content, and temporal dynamics) were extracted and compared between modalities using 2-tailed Welch t tests with Benjamini-Hochberg false discovery rate correction.

Results

App-generated videos exhibited a pronounced recovery bias: 78% (39/50) featured narrative arcs progressing from depressive states toward resolution compared with 14% (7/50) of API outputs. This divergence was reinforced across channels. App videos brightened over time (mean slope 2.90, SD 2.43 per second vs −0.18, SD 1.24 per second for the API; Cohen d =1.59; q

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

DepressionVideo RecordingHumansGenerative Artificial Intelligence

Stichwörter der Autor:innen

AIartificial intelligencecomputer visiondepressiongenerative AImental health representationtext to video

Autor:innen und Institutionen

  • Division of Digital Psychiatry, Beth Israel Deaconess Medical Center, 20 Overland St, Suite OV-202, Boston, MA, 02215, United States, 1 6176676700.
  • Digital + Media, Rhode Island School of Design, Providence, RI, United States.
  • Department of Psychiatry and Psychotherapy, Campus Benjamin Franklin, Charité - Universitätsmedizin Berlin, Berlin, Germany.
  • School of Public Health, Boston University, Boston, MA, United States.

Zitieren

APA

Flathers, M., Smith, G., Herpertz, J., Zhou, Z., & Torous, J. (2026). Depictions of Depression in Generative AI Video Models: Mixed Methods Study of OpenAI’s Sora 2. Journal of Medical Internet Research, 28, e95682-e95682. https://doi.org/10.2196/95682

Vancouver

Flathers M, Smith G, Herpertz J, Zhou Z, Torous J. Depictions of Depression in Generative AI Video Models: Mixed Methods Study of OpenAI’s Sora 2. Journal of Medical Internet Research. 2026;28:e95682-e95682. doi:10.2196/95682

BibTeX
@article{Flathers2026,
  author = {Flathers, Matthew and Smith, Griffin and Herpertz, Julian and Zhou, Zhitong and Torous, John},
  title = {Depictions of Depression in Generative AI Video Models: Mixed Methods Study of OpenAI’s Sora 2},
  journal = {Journal of Medical Internet Research},
  year = {2026},
  volume = {28},
  pages = {e95682--e95682},
  doi = {10.2196/95682}
}
Kostenlos starten →