Truace
Subscribe
60/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

Bored, lonely students were more likely to use AI chatbots problematically, a 551-student study finds

Study title: From boredom to problematic AI chatbot use: a three-wave half-longitudinal mediation study of loneliness and dual-system self-control

Evidence type
Observational study
Sample
n = 551
Sources
1 sourcePeer-reviewed
Published
2 min read
Truvace record
TRV-2026-2530cite or check it

Study title: From boredom to problematic AI chatbot use: a three-wave half-longitudinal mediation study of loneliness and dual-system self-control

A study of 551 students at two Chinese universities found that those who scored higher on boredom proneness in October 2024 reported more problematic AI chatbot use in December 2024. The direct association was β = .25, and boredom proneness was also tied to loneliness (β = .57) and stronger impulsive tendencies (β = .31).

Read the study abstract

Background: Problematic use of AI chatbots (PUAIC) is an emerging concern. Drawing on the I-PACE model and the dual-system model of self-control, this study examined the mediating roles of loneliness, the control system, and the impulsive system. Methods: A three-wave half-longitudinal mediation design was used with 551 Chinese college students at baseline. Boredom proneness was assessed at Time 1, loneliness and the control and impulsive systems at Time 2, and PUAIC at Time 3. Missing data were handled using 20…

The problem

In a three-wave study of 551 Chinese college students, higher boredom proneness was associated with more problematic AI chatbot use two months later, directly (β = .25) and through loneliness and impulsive tendencies.

In brief

Loneliness and the impulsive system each carried part of the link. Indirect effects were .073 through loneliness, .077 through impulsivity and .038 through loneliness followed by impulsivity. The total effect was .401, and the models explained 23% of the variance in problematic use.

The results are associations, not proof of cause. Baseline levels of loneliness, self-control and problematic use were not measured, so earlier problematic use may have contributed, and a reverse-order model also produced a significant pathway between impulsivity and loneliness. The authors say prevention ideas, such as work on loneliness and boredom coping, need experimental testing.

Main points

  1. Researchers recruited 551 students from two Chinese universities; 533 took part in wave 2 and 521 gave valid wave 3 responses, with complete data for 503.
  2. Boredom proneness was measured in October 2024, loneliness and self-control in November, and problematic AI chatbot use in December, with about 30 days between waves.
  3. Boredom proneness was linked to loneliness (β = .57), stronger impulsivity (β = .31), weaker control (β = -.38) and problematic chatbot use (β = .25).
  4. The total indirect effect was .156; the control-system pathways were not significant.
  5. A complete-case analysis of 503 students gave the same conclusions, with a direct effect of β = .243.
  6. Most students mainly used Doubao (68.2%) or DeepSeek (21.9%), and averaged 2.98 hours a week with chatbots.
  7. Problematic use was measured with an 11-item scale adapted from a ChatGPT-use scale, and the authors say frequent use alone is not treated as problematic.

The full story

The study used a half-longitudinal design across the fall 2024 semester. Participants were recruited by convenience sampling at two Chinese universities; 57.5% were female and the mean age was 19.02. Missing data were handled with 20 imputed datasets and PROCESS Model 81, with sex and grade level as covariates.

Within the model, loneliness was associated with the impulsive system (β = .27) and with problematic use (β = .13), and the impulsive system with problematic use (β = .24). The control system was not significantly associated with problematic use (β = .08, p = .072), and its indirect effects were not significant.

The authors suggest universities pair mental health education and digital literacy with peer activities and counselling, and that AI providers add usage reminders and referrals to psychological support. They describe these as tentative targets for future experimental evaluation, since no intervention was tested.

Sources

1 on file
  1. Australian Journal of Psychologydoi:10.1080/00049530.2026.2737094
    Peer-reviewed2026-09-27

Talk it over

Four questions to start onWho carries this cost, and what would have to change to prevent it?

Start with a question

Bored, lonely students were more likely to use AI chatbots problematically, a 551-student study findshttps://truvace.com/item/from-boredom-to-problematic-ai-chatbot-use-a-three-wave-half-longitudinal-mediat

  1. Who carries this cost, and what would have to change to prevent it?

    The problem In a three-wave study of 551 Chinese college students, higher boredom proneness was associated with more problematic AI chatbot use two months later, directly (β = .25) and through loneliness and impulsive tendencies.

    Answer in the thread

  2. How do we know? What can this kind of evidence show, and what can’t it?

    The evidence This rests on an observational study with a sample of 551, from Australian Journal of Psychology.

    Answer in the thread

  3. What does it leave open? What else would you want to know before acting on it?

    The record’s own caveat Boredom proneness, mediators and problematic use were each measured at only one wave, with no baseline controls, so causal and change-over-time conclusions are ruled out; the sample was convenience-based, from two Chinese universities, and relied on…

    Answer in the thread

  4. Have you seen AI used this way where you study, work or live? Did it go the same way?

    Answer in the thread

The questions are the same on every record. The details under them are taken from this record as written.

What readers say

More from Education