Data Driven Training: Enhancing Athletic Performance through Wearable Technologies and Artificial Intelligence

Dilican, 2025 Uluslararası Sağlık, Egzersiz ve Spor Bilimleri Dergisi, (2025) International Journal of Health, Exercise, and Sport Sciences (IJOSS) ISSN: 3023-8382 RESEARCH ARTICLE / Araştırma Makalesi Open Access/Açık Erişim IJOSS © The Author(s), 2024. Open Access. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License, which allows for unrestricted use, sharing, adaptation, distribution, and reproduction in any medium or format, provided proper credit is given to the original author(s) and the source. A link to the Creative Commons license must be included, and any changes made to the original work must be clearly indicated. Unless otherwise specified in a credit line, all images or third-party materials included in this article fall under the article’s Creative Commons license. If any material is not covered by the Creative Commons license and your intended use is not permitted by law or exceeds the permissible scope, you must obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. DataDriven Training: Enhancing Athletic Performance through Wearable Technologies and Artificial Intelligence Veri-Odaklı Antrenman: Giyilebilir Teknolojiler ve Yapay Zekâ ile Sporcu Performansının Yükseltilmesi Tunay DİLİCAN1 Abstract This review examines the increasingly prominent concept of data-driven training in contemporary sports science, discussing the role of wearable technologies and artificial intelligence in enhancing athletic performance. Based on the existing body of published research, the literature indicates substantial advancements in performance assessment, load management, and injury prevention. The large datasets generated through wearable devices can be analysed through AI algorithms, enabling the development of highly individualised training models. However, persistent challenges------such as data security, ethical standards, algorithmic bias, and financial accessibility------continue to limit the widespread adoption of these technologies. The review emphasises the need for future systems to be developed in a more transparent, accessible, and ethically grounded manner. Ultimately, data-driven approaches signal a new era in modern sport, one that extends beyond performance optimisation to encompass sustainability and a more human-centred model of athletic training. Keywords: Artificial intelligence, Performance, Training Öz Bu derleme, son yıllarda spor biliminde giderek önem kazanan veri-odaklı antrenman anlayışını incelemekte; giyilebilir teknolojiler ve yapay zekâ uygulamalarının sporcu performansının geliştirilmesindeki rolünü tartışmaktadır. Yayımlanmış çalışmalar temel alınarak, literatür taraması sonucunda performans ölçümü, yük yönetimi ve sakatlık önleme konularında önemli gelişmeler olduğu görülmüştür. Giyilebilir cihazlar aracılığıyla elde edilen büyük veri setleri, yapay zekâ algoritmalarıyla analiz edilerek bireyselleştirilmiş antrenman modellerinin oluşturulmasına imkân tanımaktadır. Bununla birlikte, veri güvenliği, etik standartlar, algoritmik önyargı ve maliyet gibi sorunlar, teknolojilerin yaygın kullanımını sınırlamaktadır. Çalışma, gelecekte bu teknolojilerin daha şeffaf, erişilebilir ve etik temellere dayalı biçimde geliştirilmesinin gerekliliğini vurgulamaktadır. Sonuç olarak, veri odaklı yaklaşımlar, modern sporun sadece performans optimizasyonu değil, aynı zamanda sürdürülebilirlik ve insan merkezli antrenman anlayışı açısından da yeni bir dönemi temsil etmektedir. Anahtar Kelimeler: Yapay zeka, Performans, Antrenman https://www.ijoss.org/Archive/issue2-volume3/ijoss-Volume2-issue3-33.pdf * Correspondence: Tunay DİLİCAN [email protected] 1Bursa Uludağ Universtiy /Orcid ID: 0000-0003-4686-6849 e-posta: [email protected] https://doi.org/10.5281/zenodo.17583376 Received / Gönderim: 08.08.2025 Accepted / Kabul: 15.09.2025 Published / Yayın: 24.10.2025 Volume 2, Issue 3, October, 2025 Cilt 2, Sayı 3, Ekim, 2025 Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 408 of 418 Introductıon In contemporary sports science, athletic performance is no longer confined to questions such as ‘‘how much an athlete trains’’ or ‘‘how strong the athlete is.’’ Instead, performance is increasingly understood through data-driven approaches that incorporate a wide range of variables, including training load, recovery dynamics, biomechanical analysis, and personalised performance strategies. In this context, wearable technologies and artificial intelligence---supported analytic systems have become critical tools for monitoring, optimising, and sustaining athletic performance. This review aims to examine how these emerging tools and methodologies are reshaping performance enhancement, what opportunities and limitations they present, and the directions in which the field may evolve in the coming years. To begin with, the rise of wearable sensor technologies within sporting environments is particularly noteworthy. Traditionally, sports scientists and coaches have relied on laboratory measurements or field observations to assess physiological indicators, such as heart rate, oxygen consumption, and muscle activation. In recent years, however, the introduction of GPS units, accelerometers, gyroscopes, biosensors, and smart textiles has enabled the real-time collection and processing of this data during training and competition. For instance, several studies have reported that wearable sensors can track heart rate, muscle activation, joint angles, and ground reaction forces, providing valuable insights into both performance and injury risk (Alzahrani & Ullah, 2024). Moreover, a comprehensive review highlighted the widespread use of wearable technologies in measurement, monitoring, training, and rehabilitation processes, while also noting key challenges related to accuracy, cost, ethics, and data security (Segura et al., 2018). The vast data generated by these systems holds significant potential when interpreted through artificial intelligence and machine learning tools. AI systems not only collect and visualise raw data but also develop predictive and prescriptive models to support load management, training optimisation, fatigue prediction, and injury-risk assessment (Valiyev & Mahmudova, 2025). In team sports, for example, AI-assisted analytics platforms monitor training loads, on-field movement patterns, and recovery statuses to reduce injury risk and enhance decision-making. This technological transformation encourages sport scientists to shift from a purely capacity-based perspective toward a culture of evidence-informed decision-making. Nevertheless, despite the considerable promise of these technologies, several important limitations remain. Technical factors such as measurement accuracy, sportspecific validity, sensor placement, and environmental conditions may introduce error or variability into the data (Olsen et al., 2025). In addition, issues related to data security, user privacy, and ethical governance have become increasingly prominent. Regulations concerning the collection, storage, and sharing of athlete data are still not fully standardised across global sporting contexts. These considerations underscore the necessity for both technological refinement and the development of robust ethical frameworks as the field continues to evolve. From an artificial intelligence perspective, however, several issues-----such as the opacity of ‘‘black-box’’ algorithms, model validity, and adaptive capacity-----remain subjects of ongoing debate. AI models are typically trained on large datasets; yet, the extent to which these models can be generalised across different athletes, performance levels, or environmental contexts continues to be an active area of investigation (Altukhi et al., 2025). In this regard, the concept of data-driven training should not be interpreted merely as the use of technology; rather, it encompasses the integration of technology Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 409 of 418 with human expertise. Instead, it encompasses a multi-stage process in which data collected from wearable sensors are transformed into meaningful training programmes, coaches and sport scientists make informed decisions, AI-assisted tools are integrated into the training cycle, and the outcomes of these processes are continuously monitored. In other words, the core components of this framework form a cyclical system that involves data collection, data analysis, application, and feedback. Through this system, training load can be adjusted according to individual athlete characteristics, adaptation to training stimuli may be accelerated, and ultimately, performance improvements can be facilitated. In conclusion, this review will first examine the influence of wearable technologies and artificial intelligence on athletic performance, followed by an exploration of how these approaches may be integrated into training processes, including practical applications, current limitations, and emerging trends. Through this comprehensive synthesis, the review aims to offer sport scientists, coaches, technology developers, and researchers an expanded and nuanced perspective on the evolving landscape of performance optimisation. Conceptual Framework To contextualise the interconnected role of data acquisition, computational analysis, and applied decision-making in modern performance science, this review adopts an integrated conceptual framework. In this model, wearable sensors operate as primary data generators, artificial intelligence systems function as analytical engines that extract meaningful patterns, and coaches translate these insights into actionable performance strategies. This triadic synergy demonstrates that performance optimisation emerges not from technology alone, but from the coordinated alignment of measurement precision, analytical intelligence, and applied sport expertise. As the integration of wearable technologies and artificial intelligence continues to expand, the relationship between data quality, model reliability, and practical applicability becomes increasingly important. The effectiveness of these systems depends not only on the accuracy of the collected data but also on the capacity of coaches, analysts, and sport scientists to interpret and contextualise these outputs within real training environments. Thus, technological innovation alone is insufficient; instead, it must be accompanied by a comprehensive understanding of athlete-specific demands, training principles, and physiological adaptation mechanisms. Establishing this synergy between technology and applied sport science creates a foundation upon which more precise, individualised, and sustainable performance strategies can be developed. Key Definitions For conceptual clarity, several key terms used throughout this review warrant explicit definition. Wearable technologies refer to sensor-based devices that continuously capture physiological, biomechanical, or environmental data during training or competition. Artificial intelligence encompasses computational systems capable of learning from data and producing predictive or prescriptive outputs. Machine learning-----a subset of AI-----is concerned with algorithms that detect underlying patterns in complex datasets and refine their predictions over time. Data-driven training denotes a methodological approach in which training decisions are informed by empirical evidence rather than intuition alone Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 410 of 418 Materials and Methods This study is a narrative review aimed at compiling and synthesising the current scientific literature on data-driven training, wearable technologies, and artificial intelligence applications in enhancing athletic performance. The purpose of the review is to identify recent research developments in this field, outline emerging trends, and highlight the opportunities and challenges encountered in practical applications. To this end, peer-reviewed articles, systematic reviews, meta-analyses, reports, and academic theses published in Turkish and English were examined. The literature search was conducted through PubMed, Scopus, Web of Science, Google Scholar, and ResearchGate databases. During the search process, the following keywords were used: “data-driven training,” “wearable technologies in sports,” “artificial intelligence and athletic performance,” “sports analytics,” “load management,” and “performance monitoring.” Inclusion Criteria  The following criteria were considered for inclusion:  Direct relevance to athletic performance, training management, or injury prevention;  Incorporation of wearable technologies or AI-based analytical methods;  Publication in either English or Turkish. Research Design This study employed a narrative (descriptive) review design. Unlike systematic reviews, the aim of narrative reviews is not to test a specific hypothesis but rather to summarise, interpret, and synthesise existing findings within the literature in a thematic manner. The research design consisted of three main stages: 1. Literature Search A systematic search of the aforementioned databases was conducted, and the titles, abstracts, and full texts of the retrieved studies were reviewed. 2. Data Extraction and Classification  Eligible studies were coded based on:  type of research (e.g., experimental, observational, review),  type of technology used (e.g., GPS, accelerometer, artificial intelligence, machine learning),  sport discipline (e.g., football, running, swimming),  Moreover, reported performance outcomes (e.g., endurance, speed, load management, injury incidence). 3. Thematic Synthesis The studies were synthesised under the following thematic categories:  Technological applications, Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 411 of 418  Data analytics and artificial intelligence approaches,  Performance monitoring and load management,  Practical limitations in applied settings,  And ethical considerations. This research design enabled the review to encompass a broad range of literature while also identifying emerging patterns and directions that may guide future research in this rapidly developing field. Data Analysis In this review, the data extracted from the included studies were evaluated using a qualitative content analysis approach. Rather than applying quantitative statistical procedures, the key concepts, methodological approaches, reported outcomes, and research trends highlighted in each study were systematically compared and analysed. Each study was examined in relation to its primary variables, including:  the type of wearable device used (e.g., GPS, heart-rate monitors, EMG sensors),  the performance parameters measured (physiological, biomechanical, technical–tactical),  the artificial intelligence or machine learning methods implemented,  Moreover, the reported outcomes are related to performance enhancement or risk reduction. These codes were subsequently merged into broader thematic categories, from which overarching trends in the field were identified. The thematic analysis indicated three major patterns:  real-time data tracking and personalised training approaches,  machine learning–based performance prediction models,  Moreover, considerations related to data privacy, ethics, and practical limitations must also be taken into account.  This analytical process provided the foundation for the detailed discussion of each theme presented in the subsequent sections of the review. Findings Recent technological advancements in sports science—particularly the emergence of wearable sensor technologies and AI-supported analytical methods—have positioned these tools as critical instruments for monitoring, optimising, and sustaining athletic performance. In this section, the rise and applications of wearable technologies in sports will be examined first, followed by an exploration of how artificial intelligence and machine learning approaches contribute to this domain. Finally, the combined opportunities and challenges that arise from the integration of both approaches will be analysed. Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 412 of 418 Wearable Technologies and Athletic Performance Wearable sensors include devices such as GPS units, accelerometers, gyroscopes, heart-rate monitors, and intelligent textiles. Through these instruments, athletes’ physiological variables (e.g., heart rate, oxygen consumption), biomechanical parameters (e.g., joint angles, muscle activation), and load-related metrics (e.g., ground reaction force, acceleration) can now be monitored in real time or with minimal delay during both training sessions and competitive environments. For example, a systematic review investigating wearable devices in sports such as field hockey found notable variability in measurement accuracy across devices, although overall utilisation has been steadily increasing (Latino & Tafuri, 2024). From the perspective of achieving small yet cumulatively meaningful performance gains, the role of wearable technologies has been widely discussed in the literature. One study categorised wearable devices into three primary groups: location-based wearables, biometric wearables, and performance-monitoring wearables. This classification is valuable for coaches and sport scientists, as it clarifies which types of sensor data are associated with specific performance outcomes (Migliaccio et al., 2024). The advantages offered by wearable sensors in training load management have also been highlighted. Research has shown that variables such as total loading, number of direction changes, and acute-to-chronic workload ratios are frequently assessed in sports environments where these technologies are employed (Rebelo et al., 2023). According to recent reports, wearable technologies have become a growing trend within the fitness and sports science sectors, with their global market size projected to reach $ 186 billion by 2030 (Doherty et al., 2024). Despite this rapid expansion, the literature also documents several technical and practical limitations, including issues related to measurement accuracy, sport-specific validity, sensor placement, environmental conditions, data connectivity, and the interpretation of sensor outputs. Studies have further demonstrated substantial inter-individual variability in sensorderived data, emphasising the need for personalised approaches (Baldassarri et al., 2023). Although wearable technologies are widely recognised as powerful tools for athlete monitoring and training management, their practical use depends on careful consideration of factors such as validation, standardisation, and user compliance. These elements remain essential to ensuring that wearable systems yield reliable and actionable performance insights. Artificial Intelligence and Machine Learning Approaches The large volumes of data generated by wearable sensors are not inherently meaningful on their own; this is where artificial intelligence (AI) and machine learning (ML) techniques become essential. These technologies enable raw data to be transformed into actionable insights, allowing for the development of proactive strategies in areas such as load management, fatigue prediction, and injury-risk assessment. For instance, a comprehensive review reported that AIand ML-based approaches have made significant advancements in sports science, particularly in injury forecasting, performance analysis, and personalised training programme design (Reis et al., 2024). Another study examined the applications of AI within sporting contexts, including Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 413 of 418 the challenges associated with ethical considerations and data quality, as well as emerging future directions (Zhou et al., 2025). Specifically, ML algorithms are increasingly used to analyse movement patterns, enabling early detection of fatigue or heightened injury risk. In one investigation, high-dimensional data derived from wearable sensors were processed using AI techniques to identify biomechanical risk factors, demonstrating the potential practical value of such analyses (Musat et al., 2024). However, several barriers still hinder the broader adoption of AI in sport. These include the opacity of “black-box” algorithms, variability in data quality, and ethical concerns such as data privacy and algorithmic bias. Therefore, AIand ML-supported performance analytics require more than the mere implementation of advanced technologies; they also necessitate ensuring that these systems are reliable, valid, ethically governed, and compatible with end-user needs. This highlights the importance of integrating technological sophistication with robust methodological and ethical frameworks to maximise the practical utility of AI in high-performance sport. Integration of Wearable Technologies and Artificial Intelligence: Performance Enhancement and Sustainability In recent years, the integration of artificial intelligence and wearable technologies has emerged as a transformative approach within sports science, offering substantial potential for enhancing performance and improving training efficiency. Wearable devices—such as GPS-based systems, accelerometers, EMG sensors, and heart-rate monitors—collect real-time physiological and biomechanical data, which are subsequently analysed through AI algorithms to generate personalised performance models (Li & Washington, 2024). This integration enables coaches to adopt data-driven decision-making processes, facilitating more precise management of key performance parameters. One of the most significant advantages of AI-supported wearable systems is their ability to convert complex datasets into meaningful insights. Machine learning algorithms, in particular, can identify patterns within large volumes of data to predict variables such as training load, injury risk, or recovery status, and can provide personalised recommendations accordingly (An, 2025). For example, studies conducted in professional football and athletics have demonstrated that these systems can achieve accuracy rates exceeding 85% in predicting training loads. However, existing research also highlights several limitations concerning the sustainable use of these technologies. Issues such as data security, economic accessibility, and the lack of long-term validation for many devices represent prominent challenges. Studies have noted that ethical considerations and data bias in AI-supported sports technologies remain partly unresolved, and that algorithmic bias may introduce errors into performance assessments (Olyasanab & Annabestani, 2024). Moreover, the sustainability of these systems depends not only on their technical performance but also on user compliance and the robustness of data-privacy protocols. A further concern relates to economic feasibility. Many of the current AI-integrated wearable systems require high-cost hardware and continuous cloud-based dataprocessing infrastructures. This creates accessibility barriers, particularly for youth Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 414 of 418 athletes, amateur leagues, and financially constrained clubs. Consequently, future research may benefit from prioritising the development of more inclusive, cost-effective, and transparent systems that ensure data security while remaining practical for a broader range of users. Explicit Research Gaps Despite promising advancements in wearable technologies and AI-driven analytics, several persistent research gaps remain. Many studies rely on narrow samples, shortterm observations, and limited data diversity, which restrict the generalisability of their results. Additionally, few investigations examine how these technologies perform under varying environmental conditions, in different competitive levels, or across diverse sport disciplines. This lack of methodological breadth limits our understanding of long-term adaptation, algorithmic robustness, and real-world applicability. Gaps in the Literature and Future Directions Although the existing literature highlights the considerable potential of integrating wearable technologies and artificial intelligence for enhancing athletic performance, most studies remain at the pilot stage or are limited to specific sports disciplines (Sampedro, 2023). Research examining the ethical dimensions of these technologies indicates that issues such as data ownership and informed consent from athletes are frequently overlooked in practice, which may jeopardise athletes’ long-term psychological and legal rights. Additionally, the diversity and quality of data used in AIbased performance analysis remain limited. A comprehensive review noted that insufficient data diversity persists as a significant challenge in many sports-related AI systems, ultimately restricting the generalisability of the algorithms. Consequently, future research may benefit from the adoption of more heterogeneous samples and multimodal data sources—such as sensor-based measurements, video analysis, and environmental parameters (Radanliev, 2025). Methodological gaps also exist concerning long-term performance monitoring and sustainable adaptation. Many technological studies conducted on athletes are restricted to short-term experimental designs, despite evidence that training adaptation unfolds over extended periods, often spanning several years (Carron & Dalbo, 2025). This creates a methodological barrier to a comprehensive understanding of the long-term impact of wearable technologies on performance development. Looking toward future directions, AI-assisted systems are expected to play an increasingly prominent role in personalised training and predictive health monitoring. In particular, the integration of real-time sensor data with cloud-based AI systems may enable athletes to optimise training load more effectively and anticipate injury risk with greater accuracy. At the same time, the establishment of international standards concerning ethics, privacy, and data fairness will likely serve as a critical determinant of the future adoption and scalability of these technologies. Structured Directions for Future Research Future research may benefit from adopting more rigorous and multidimensional Dilican, 2025. International Journal of Health, Exercise, and Sport Sciences Vol 2, issue 3, October 2025 Page 415 of 418 approaches. Longitudinal cohort studies could clarify how athletes adapt to AI-assisted training over extended periods, while multimodal datasets—integrating sensor data, video analytics, and contextual variables—may enhance ecological validity. Technological developments in explainable AI, edge computing, and decentralised data governance could improve transparency and security. Ethically, internationally aligned standards concerning consent, privacy, and algorithmic fairness will be crucial to ensuring responsible deployment. Human–AI Interaction in Performance Decision-Making Although AI systems provide sophisticated predictive capabilities, decision-making in high-performance environments must remain fundamentally human-centred. Coaches interpret algorithmic outputs through contextual knowledge, experiential judgement, and athlete-specific nuances that computational models cannot fully replicate. Thus, AI should serve not as a replacement for expert insight but as an augmentative tool that enhances a coach’s ability to detect subtle performance deviations, manage uncertainty, and design personalised training strategies. Limitations of This Review This review is subject to certain limitations. The literature search was restricted to specific databases and to studies published in English and Turkish, which may have excluded relevant research in other languages. As a narrative review, the synthesis does not follow the strict methodological protocols of systematic reviews, introducing the possibility of selection bias. Furthermore, heterogeneity across study designs limits direct comparability. These limitations highlight the need for more systematic and longitudinal investigations. Conclusion This review examined contemporary trends in the use of wearable technologies and artificial intelligence applications for monitoring, evaluating, and enhancing athletic performance. The literature suggests that data-driven approaches have significantly contributed to performance gains in both professional and amateur sports. In particular, the accuracy of AI-based systems in real-time data collection, load management, and injury-risk prediction has steadily improved. Wearable devices possess the capacity to measure not only physical performance indicators but also physiological and psychological feedback, thereby offering notable potential for personalising training processes. Nevertheless, issues such as data security, ethical responsibility, algorithmic bias, and user privacy remain unresolved. These concerns underscore the need for robust ethical and legal frameworks to accompany technological advancements. Additionally, the existing literature reveals that many studies rely on short-term pilot applications and are limited in scope to specific sport disciplines. This presents a methodological challenge for assessing the long-term effects of AI and wearable systems on athletic development. Practical Implications for Applied Settings