Under a common activity target and monitoring environment, intelligent-assistant support was associated with small shifts in the composition of extracurricular physical activity, particularly towards running, but not with greater overall participation or consistently more favourable anthropometric and fitness outcomes.
Key Findings
Results
Intelligent-assistant support was associated with slightly fewer total extracurricular activity sessions than usual self-scheduling.
Semester-long non-randomised comparative study of 6,530 undergraduates (3,258 intelligent-assistant; 3,272 usual self-scheduling).
The assistant group had 0.58 fewer total activity sessions than the usual self-scheduling group.
Both groups faced the same institutional requirement of 45 extracurricular sessions.
Group differences were tested with Mann-Whitney U tests and covariate-adjusted linear regression.
Results
Intelligent-assistant support was associated with a higher proportion of sessions involving running.
The assistant group had a 2.13-percentage-point greater proportion of sessions involving running.
The authors describe this as a small shift in the composition of activity, particularly towards running.
Results
Intelligent-assistant support was associated with a longer mean distance per running session.
Mean running distance per session was 0.061 km greater in the assistant group.
This was estimated with covariate-adjusted linear regression.
Results
Vital capacity was higher in the intelligent-assistant group.
Vital capacity was 41.4 mL higher than in the usual self-scheduling group.
This was an end-of-semester fitness outcome.
Results
Endurance-run time was lower (better) with intelligent-assistant support among male students only; no between-group difference was seen among female students.
Endurance-run time was 2.68 s lower among male students in the assistant group.
No between-group difference was observed among female students.
Endurance-run time was sex-specific (different run distances for each sex).
Results
Body mass index was higher in the intelligent-assistant group.
BMI was 0.36 kg/m2 higher than in the usual self-scheduling group.
The authors conclude outcomes were not consistently more favourable with the assistant.
Conclusions
The intelligent-assistant did not increase overall participation or yield consistently better anthropometric and fitness outcomes.
The assistant integrated timetable-informed scheduling, adaptively timed reminders based on evolving exercise habits, and immediate feedback.
Both approaches operated within existing physical education course sections under a shared activity target and monitoring environment.
The design was non-randomised, so findings are described as associations.
What This Means
This study compared two groups of university students over one semester. About 3,250 used an intelligent assistant that built exercise schedules around their class timetables, sent reminders timed to their habits, and gave immediate feedback. About 3,270 scheduled their own exercise as usual. Everyone had to complete 45 extracurricular activity sessions. Students were not randomly assigned to groups.
The assistant did not get students to exercise more. Its users logged slightly fewer sessions (0.58 fewer). They did shift somewhat toward running, with a 2.13-percentage-point higher share of sessions involving running and slightly longer runs (0.061 km more per run). Lung capacity was a little higher (41.4 mL), and male students had slightly faster endurance-run times (2.68 s), but female students showed no difference. BMI was slightly higher (0.36 kg/m2) in the assistant group.
This research suggests that digital support tools like this may change which activities students choose more than how much they do, at least when a mandatory activity target already exists. The differences were small, and because students were not randomly assigned, the study cannot show that the assistant caused them. Benefits for fitness and body measures were not consistent.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Li X, Jiang Y, Wang Y, Wang Y, Liu Y. (2026). Intelligent-assistant support versus usual self-scheduling for extracurricular physical activity among university students: a comparative evaluation.. Frontiers in public health. https://doi.org/10.3389/fpubh.2026.1945553