A 10-week Large Language Model (LLM)-generated versus human-made volleyball training program on the jumping performance of collegiate volleyball athletes

Abstract

The study compared the effectiveness of AI-generated and human-generated volleyball training programs in enhancing vertical and horizontal jumping performance among 43 student-athletes. Participants were divided into an AI-generated volleyball training group (23 participants) and a human-made training group (20 participants), each following a 10-week, three-session-per-week program. Vertical and horizontal jumping performances were measured before and after the training programs using vertical and standing broad jump tests. Paired t-tests were used for within-group comparisons, and independent t-tests analyzed post-test differences. The AI-generated program improved right unilateral vertical and horizontal jumps but did not significantly affect bilateral or left unilateral jumps. Female athletes showed improvements in bilateral and right unilateral vertical jumps. The human-made program enhanced bilateral and right unilateral vertical jumps, especially in females, but did not significantly affect horizontal jumping performance. The study concluded that AI-generated programs have potential but have yet to be a substitute for professionally designed training programs. Further research is recommended to expand the study to include volleyball-specific skills.