Worry that artificial intelligence could eventually replace parts of a person’s work was followed by a small rise in cynicism toward studying among medical and nursing students in a 10-day diary study conducted in China.
The observational study followed 751 undergraduates at one university in northern China. Students completed evening surveys on 10 weekdays across two weeks in May and June 2025. The analysis examined whether a student’s higher-than-usual worry on one day was associated with changes in feelings reported at the next diary entry.
What the study found
Students rated their agreement with the statement, “I worry that what I learned today and what I did today may be replaced by AI in the future,” on a five-point scale. The average rating was 2.69. About 29% of the variation in worry reflected day-to-day changes within the same student, rather than stable differences between students.
Higher-than-usual AI replacement worry was followed by lower enthusiasm for studying at the next diary entry and was associated with more cynicism toward studying. The association with cynicism was small: the standardized coefficient was 0.068, with a 95% credible interval from 0.016 to 0.119.
In turn, higher cynicism was associated with more anxiety at the following diary entry. That coefficient was 0.067, with a 95% credible interval from 0.019 to 0.112.
However, the researchers did not confirm that cynicism carried AI worry forward into anxiety. The indirect effect’s credible interval included zero in all six models. The direct association between AI worry and anxiety two diary entries later was also not supported when the intermediate steps were removed from the analysis.
What was not linked to AI worry
AI replacement worry was not associated with next-day exhaustion or with the response to whether students had learned anything interesting that day.
The study did find that a more anxious day was followed by higher scores for not learning anything interesting, more exhaustion and more cynicism. But the reverse association between feeling that nothing interesting was learned and later AI worry was not consistent across the analyses.
These findings suggest that students’ feelings may move together across days, but they do not show that AI worry causes burnout or anxiety. Exams, workload, sleep and exposure to AI-related news could also have affected students’ responses over different timescales.
Why the design matters
The researchers used dynamic structural equation modeling. Rather than comparing only students who generally reported more or less worry, the approach compared each student with their own average and examined whether higher-than-usual scores came before changes at the next diary.
That design can provide a closer look at short-term patterns than a single survey. It still cannot establish causation. The study measured brief fluctuations over about two weeks, not lasting burnout or clinical anxiety disorders.
Students submitted 4,625 of 7,510 planned diaries, or 61.6%. They averaged about six diaries each, and only 560 students provided a complete three-day run. The researchers reported that the two cynicism links remained supported when the Friday-to-Monday gap was handled using calendar spacing and in a stricter sample of 695 students with at least three valid diaries.
A wider concern in health education
The study sits within broader questions about how health-professions students are preparing for AI-enabled workplaces. A 2024 survey of 542 medical students reported high AI anxiety and moderate readiness for AI; students who felt more ready reported less anxiety about it.
Career preferences may also be shifting in some settings. In a survey of 463 medical students in the United States, 21.4% ranked radiology first when AI was not considered, compared with 17.7% when it was considered.
Burnout is an established concern in nursing education, although the figures come from different research. A 2024 meta-analysis of 34 studies involving 9,554 nursing students estimated academic burnout prevalence at 35%, with the highest reported rate, 58%, in Asia. Those estimates cannot be used to calculate the prevalence of burnout in the diary-study group.
What may help researchers and educators understand next
The study’s researchers suggested that responses to AI-related worry in health education could include professional meaning, support for competence, and clearer discussion of how people and AI may work together alongside technical AI skills. That approach has not been tested in a trial, so its effects are unknown.
For readers, the main message is narrower: day-to-day concern about AI replacement was associated with small, short-term changes in how students viewed studying. The evidence does not show that AI worry produces burnout, anxiety disorders or long-term academic effects.
AI tools were used to assist with the preparation of this article under our editorial standards.
Leave a Reply