Key result
AI models show substantial potential for football talent identification despite data quality and validation limits.
Why the study?
Existing evidence on AI in football talent identification was dispersed, and a systematic synthesis of application methods and effectiveness across different identification stages was lacking.
Systematic Review (n=20)
AI shows substantial potential in football talent identification, but current research is still transitioning from methodological exploration to systematic application and requires better datasets and external validation.
AI shows value across talent ID stages; supports expanded use while highlighting needs for interpretability and real-world validation.
Background The rapid development of artificial intelligence (AI) in sports has driven a shift in football talent identification from traditional experience-based judgment to data-driven decision-making. However, existing evidence on the application of AI in talent identification is dispersed, and a systematic synthesis of its application methods and effectiveness across different stages of the identification process is lacking. Objective To systematically review the application of AI in football talent identification, compare sample characteristics, model types, data sources, and application contexts across different AI techniques, and summarize their advantages, limitations, and future directions. Methods This systematic review followed PRISMA guidelines and was registered with PROSPERO. Searches were conducted in Web of Science, Scopus, SportDiscus, and PubMed to identify eligible studies, which were assessed for methodological quality. Information regarding sample characteristics, AI technique type, input features, model performance, and application context was extracted and synthesized. Results A total of 20 studies were included, covering youth and professional male players. Traditional machine learning models were the most frequently used, whereas deep learning and hybrid models were less common. AI demonstrated potential in three main areas: developmental potential identification, performance and role identification, and player value and recruitment decision-making. Although most studies adopted multidimensional feature sets, limitations remain regarding data quality, model interpretability, methodological rigor, and external validity. Conclusion AI has shown substantial value in football talent identification, but current research is still transitioning from methodological exploration to systematic application. Future work should focus on developing more representative datasets, promoting cross-institutional data sharing, enhancing model transparency and interpretability, and validating AI models in real-world football contexts to ensure their safe, reliable, and effective application in talent identification practice. Systematic Review Registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420251196817 , PROSPERO: CRD420251196817.
No takes yet. Share an insight, caveat, or question.
Zhou et al. (2026) conducted a systematic review in Football talent identification (n=20). Artificial intelligence models was evaluated. Artificial intelligence models demonstrated substantial potential in football talent identification across developmental potential, performance, and player value, though current research is limited by data quality and external validation.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: