ARTIFICIAL INTELLIGENCE IN SPORTS PERFORMANCE AND COACHING: CURRENT APPLICATIONS AND EMERGING CHALLENGES

Abstract

Artificial intelligence is increasingly used to transform athlete data into predictions, classifications, and decision-support outputs for coaches and sport scientists. This review examines current applications of machine learning, deep learning, computer vision, and generative artificial intelligence in sports performance and coaching. A structured narrative approach was used to synthesize peer-reviewed research published mainly between 2018 and 2026, with emphasis on athlete monitoring, training individualization, technical and tactical analysis, injury-risk estimation, talent identification, and coach learning. The evidence shows that artificial intelligence can automate movement recognition, combine high-frequency wearable and tracking data, identify non-linear performance patterns, and improve the speed of video and match analysis. However, predictive accuracy reported in experimental studies does not automatically translate into effective coaching decisions. Small samples, class imbalance, inconsistent outcome definitions, limited external validation, data leakage, poor calibration, and low model explainability remain frequent problems. Generative tools can support planning, report preparation, and reflective practice, but their outputs require verification because they may omit context or generate inaccurate recommendations. The most appropriate role of artificial intelligence is therefore decision support rather than autonomous decision-making. Practical implementation should begin with a clear coaching question, validated data, interpretable outputs, and continuous human oversight. Future research should prioritize multicenter validation, female and youth athlete representation, sport-specific benchmarks, transparent reporting, and prospective studies that measure whether AI-supported decisions improve performance, health, and coaching efficiency.