Skip to main content
Preprint Credible — Major labs, established NGOs, reputable named-author preprints

Everyday use of AI for emotion regulation depends on the person and the situation

Thirteen-day ecological momentary assessment study in which 3,048 Chinese university students aged 18 to 29 reported emotional situations, emotion-regulation strategies and affect four times a day, yielding 134,275 prompts. The study estimates how often AI is used for confiding, advice or emotional support in daily life, which students and situations predict that use, and whether AI use is followed by changes in momentary affect.

Publisher

PsyArXiv (OSF); Boston University; Purdue University; Brown University; Huainan Union University; Wuhan University; Harvard University

Published

20 Sept 2026

Added

today

DOI

Key Findings

  • AI use for confiding, advice or emotional support was reported on 2.0% of 134,275 prompts and by 17.3% of participants at least once over 13 days
  • Person level: use concentrated among students with greater lifetime AI use (odds ratios up to 3.12)
  • Situation level: AI use was more likely in digital, academic and interpersonal situations (odds ratios 1.22 to 2.56)
  • AI use did not predict next-prompt affect overall, but was followed by lower positive affect when neither stressors nor positive events had occurred; the association weakened as more events co-occurred

Methodology Notes

Ecological momentary assessment with four daily prompts; analytic window days 4 to 16 (13 days); 81 participants who completed only the first three days excluded; mean age 19.33 (SD 1.00). Ethics approval Wuhan University; administered in Chinese; not preregistered; data available on request; analysis code on OSF. The deposit states the manuscript is under review and references a Nature Portfolio reporting summary. Single-country student sample; self-report. No DOI on the OSF record at fetch time. Verified through the OSF API record (published 20 September 2026, moderation state accepted) and the deposited manuscript read by the sweep.

Authors

Li, Simon M., Lin, Junkang, Sun, Shufang, Xu, Shanshan, Yu, Bin, Chen, Mark S.

Tags

emaemotion-regulationchinauniversity-studentsbase-ratepsyarxiv

Cite This

APA

Li, Simon M. et al. (2026). Everyday use of AI for emotion regulation depends on the person and the situation. PsyArXiv (OSF); Boston University; Purdue University; Brown University; Huainan Union University; Wuhan University; Harvard University. https://osf.io/preprints/psyarxiv/3gyab_v1