Integration of an AI-Powered Personalized Exercise Program for Enhancing Physical Fitness of Female Amateur Athletes

Abstract

In resource-constrained higher-education settings, amateur athletes are often not individually prescribed exercise, limiting fitness gains through training. Recent artificial intelligence (AI) advancements provide a scalable option for generating personalized programs. However, evidence on their effectiveness in young female amateur athletes from South Asia remains limited. This study evaluated the application of an eight-week Artificial Intelligence based individualized exercise program on physical fitness of female amateur athletes of Bachelor of Science (Sports Sciences) program of Lahore, Pakistan. A group of 45 female amateur athletes (Mage = 22.78 years, SD = 2.74) were selected for a pre experimental one group pre-test post-test experiment. Baseline anthropometric data (body weight, body mass index) and physical fitness measurements, the 20m shuttle run test for cardiorespiratory fitness, modified push-up test for upper body muscular endurance, plank hold test for core endurance and sit and reach test for flexibility were collected through a structured baseline assessment sheet. A custom AI tool developed weekly personalized exercise programs based on participants' reported progress and adapted from ACSM (2022) guidelines. Same post-test measurements were conducted after 8 weeks. The data analysis was performed by means of paired-sample t-tests, Cohen's d effect sizes, and 95% confidence interval through IBM SPSS Statistics (v.26). There were significant pre-to-post improvements for all six outcome variables (p < .001), and all effect sizes were large (|d| > 0.80). The largest effect size phenomenon was observed for plank hold time (d = 3.36) followed by sit-and-reach flexibility (d = 2.61), modified push-ups (d = 2.42), body weight (d = −1.85), body mass index (d = −1.81) and estimated VO₂max (d = 1.27). A study showed that a personalized exercise program powered by Artificial Intelligence delivered significant results in several physical fitness components of female amateur athletes. The results of the study demonstrate that low cost survey driven AI personalization is feasible and effective in the Pakistani higher education context and paves the way for larger controlled trials.