Artificial intelligence in sport psychology: Implications for the identification and development of talent

Artificial Intelligence (AI), defined as computational systems capable of executing tasks requiring human-like intelligence, such as reasoning, planning, learning, perception, and interaction, represents a transformative paradigm within sport psychology. Machine Learning, a key subdomain of AI that enables pattern recognition and predictive modeling from data, has demonstrated significant efficacy in performance prediction and talent identification across different sports, as evidenced by Baker et al. (2025) in their review. However, the implementation of these technologies faces significant hurdles, critically concerning data availability, quality, ownership, and the inherent challenges of labeling for supervised learning. Longitudinal studies are further complicated by missing data and dataset imbalance, particularly concerning outcomes such as the rare development of talent into top-tier athletes, which can lead to biased models with poor generalizability. Standardized development frameworks, such as the Software Development Life Cycle, are advocated to mitigate these risks. Future directions emphasize human-machine interaction, such as conversational AI for decision support and digital twins for athlete modeling. Nevertheless, psychological and subjective factors, like coach judgment, athlete attitudes are critical yet currently underrepresented. AI will be playing a crucial role in enhancing communication between domain experts (coaches, scouts) and complex "black-box" systems, building trust for collaborative talent development. Future advancements should prioritize the integration of psychological constructs through idiosyncratic and idiographic approaches, while exploring deep learning for safeguarding applications. [Abstract copyright: Copyright © 2025 Elsevier Ltd. All rights reserved.]