The application of different technologies to enhance sports training has been steadily progressing over the last thirty years. From football, tennis and rugby, to gymnastics and trampolining, different technologies are now being used for the collection of data to inform athlete skill development and performance (Jaswal, Stergiou and Katz, 2022).

Fitness trackers, embedded sensors and smart watches can be used to monitor aspects of athlete training, including heart rate, speed and agility. Data analytics, such as wearable devices, can be used to monitor and optimise training programmes and virtual augmented reality has the capacity to support athlete skill practice in different scenarios. These technological developments have accelerated with advances in Artificial Intelligence (AI) and machine learning to support data analytics which can track body movements and performance to deliver vast amounts of data for athletes, sports coaches and judges (Bridgeman and Giraldez-Hayes, 2024).   

  

Many of the technologies used in sports are often assumed to work directly with AI when, as Couzens (n.d.) states, they predominantly rely on algorithms, indicating that the role AI plays in sports is not always clear.  

So, what do we mean when we talk about AI in sports and what can it do?

‘A branch of computer science that aims to create machines and systems that can perform tasks that normally require human intelligence’ (FasterCapital, 2025).

AI in sports essentially refers to systems or technologies that carry out tasks normally performed by a human. AI has infiltrated many different areas of sports. From sport commentary, recruitment and judging, to coaching and athlete training.

Movement patterns in trampolining and gymnastics, for example, have become much faster and more complex, reflecting performer skill development. Such complexity makes coach assessment of performance very difficult. Data driven solutions, including the application of AI support is emerging as a potential solution, using Explainable artificial intelligence (XAI) to create a system that can classify complex movements (Woltmann et al., 2023).

AI has the capacity to analyse large sets of data, such as analysing football players to allow automatic identification of dynamic attack formations which can support ‘tactical training’ (Sperlich, Duking, Leppich and Holmberg, 2023, p. 2). AI also affords opportunities for the identification of athlete talent, predicting risk of future injury, as well as the reduction of human error in sports judging and coaching (Bharadwaj, 2024).

Human-based judging is assumed to be susceptible to human error due to variety of factors including judge fatigue and internal biases, and can also be influenced by judges’ values, their experience and training as well as cognitive and sensory limitations (Mazurova et al. 2021). To remedy these errors, Mazurova et al. (2021) explain that international tech companies such as Fujitsu and FIG are collaborating to develop an AI-powered judging system with the potential to eliminate human biases. FasterCapital (2025) argue that, through using AI, large amounts of data can be shared, monitored and verified and so is seen as more transparent, more objective and potentially much fairer than human based decision making which relies on instincts and in some sports is based on individual artistic interpretation.   

The role of the sports coach entails assessing vast quantities of athlete data pertaining to their health, capabilities and performance. The process of analysing data collected from various devices can take a huge amount of time, especially when dealing with high-level athletes that require a large amount of data collection. This process is very labour-intensive and can mostly only be done well at the highest levels of sport where full-time data scientists can be employed to support the coach (Couzens, n.d.). 

It is important to recognise that for many coaches across different sports, the data they receive from AI still requires some level of interpretation. Furthermore, the accuracy of AI and its value in supporting coaches and athletes very much depends on the type and quality of data that is inputted into the AI system.  

Although AI has the potential to make more accurate decisions based on data and minimisation of human error, technical feedback from AI may be confusing or difficult to interpret by some coaches, leading potentially to mistrust, misuse or confusion (FasterCapital, 2025). Ongoing debates therefore question how far AI can be trusted to replace or complement the human coaching role.

What do coaches and athletes think about the use of AI in sport?

A recent small-scale study carried out by Kofi Elhaggagi – a research intern at the Children’s Research centre – examined perceptions of AI as a coaching system or aid in trampolining. The primary distinguishing feature between an AI system and aid are its dependency on a human coach’s presence. The AI system is described as essentially being an independent AI coach, whereas an AI aid is a tool used by coaches to enhance their own coaching.

Semi-structured interviews and an open-ended survey with athletes, coaches and academics revealed insights which support many of the positive claims associated with AI but also raised questions about safety, ethical practices and the importance of human relationships in coaching.

What can AI do?

Despite 66 per cent of the research participants expressing reservations about using AI, many of them talked about the significant opportunities that it presented. This included the capacity for AI to predict and prevent athlete injury, and the capacity to support athlete skill development through performance analysis. One coach for example, describes the potential in AI:

‘To see alignment issues that are hard to determine with the naked eye … such as feet position … or weight distribution on take-off … and as coaches you take a guess at what went wrong based on your knowledge …’

Many of the coaches also described the capacity of AI aids to carry out routine tasks thereby freeing up more of their time to focus elsewhere, thus reducing the workload, stress and pressure of the role. It was also suggested that this might have the knock-on effect of making the coaching role appear more attractive to others – thus recruiting more people into the profession.

‘AI can be used for session planning and annual planning’.

Coaches also talked about how using AI might provide opportunities to gather data more quickly which can support their decision making, thereby enhancing confidence in their role.

The focus on reducing bias in coaching was also raised suggesting that AI may provide ‘more accurate technical feedback … and maybe less bias’.

Despite focusing on the potential associated with using AI in trampolining, many of the coaches drew attention to how increasing use of AI poses risks to performer safety and safeguarding practices.

Safety and safeguarding

‘I do not think AI should make decisions on progressing a skill … due to safety and the ability for humans to include the mental aspect of the sport’.

Many of the coaches communicated concerns regarding the potential ‘high risks’ posed by AI and apprehensions regarding safety, with performers indicating how they ‘would feel unsafe without human supervision’. 

Many of the coaches queried how safe an AI coach would be in trampolining when a ‘key element to coaching is the physical spotting and supporting’ which they suggested ‘AI would not be effective at’.

AI applications in trampolining are currently used for explaining and classifying performance as the technology to prescribe solutions and suggest performance adaptations is still under development. To gather information needed by an AI system to make decisions, monitoring devices such as cameras and microphones are required to record scenarios, and then this data is sent to the AI system. This can raise privacy concerns between coaches and performers who may be less inclined to open up, knowing that a recording is taking place (Bridgeman and Giraldez-Hayes, 2024). Sperlich et al. (2023) state that robust data protection is vital to ensure AI does not use or share personal athlete data, which raises implications for safeguarding practices within sports that can ensure performer safety and protect their personal data.

While in many respects AI has the potential to enhance different parts of the coaching role the capacity of AI to simulate the relational, psychological aspects of coaching that foster effective athlete-coach relationships has been called into question.

The psychology of coaching

‘A coach needs to be able to ‘read’ the gymnast’.

The coach-performer relationship was frequently stated by the coaches and performers as being imperative for maintaining the enjoyment of the sport and for successful training relationships.

‘They need a coach to be there, comforting them, managing their worry and building their confidence’.

The capacity to understand, often intuitively, and ‘read’ the performer was highlighted by many of the research participants as a central feature of an effective coach-performer relationship.

A human coach would likely be better at reading their gymnast’s body language and adapt their coaching to make the gymnast feel more comfortable.

Many also expressed concerns that increased AI usage may remove the individualism and artistry of the sport, which is one of the main attractions for spectators and performers. The ability to interpret emotions was repeatedly mentioned as a deficit of an AI coach, stating how it is crucial for coaches to fully understand whether a performer is confident and ready to attempt new exercises.

‘Coaching requires a human touch to ensure that competitors are being taught with a style that suits them’.

This raises questions as to whether aspects of the coaching role can be reduced to technical skill-based support, as one coach argues:

‘AI would say something technical … this is not what the child needs to hear’.

Bachikova (2024) argues that while AI excels in solving problems based on patterns in data, coaching is fundamentally a human-centred process that is underpinned by skills in empathy, deep understanding and the capacity to relate to and respond to complex, nuanced situations. These are skills that AI does not pose. Furthermore, Bachikova suggests that coaching is much more than problem-solving and more about human growth and self-awareness which require relational connections and understanding, which only human coaches can provide.

The future of AI in coaching

Rather than see AI as replacing the human aspects of the coaching role, many of the coaches emphasised the value of using AI as an aid – as something that could support and enhance their role. An AI coach in its entirety may not be feasible at this point in time, due to the limitations in technology to mimic human relationships and to interpret and act on emotions, as well as its inability to carry out physical actions such as catching, matting and spotting. More research is required to explore how coaches and performers can draw on their experience to feed into AI development as aids which can support the coaching-performer relationship. Furthermore, research which can test out AI and develop beyond explanatory technology to prescriptive technology feels timely.

There can be no denying, as Taylor-Price (2024) claims, that AI in sports is very much here to stay but closer analysis raises questions as to how safe it is for certain sports and how far it can go in simulating or complementing the relational and very human aspects of sports that rely on reciprocal relationships and intuitive coaching.

References

Bachkirova, T. (2024) ‘Why coaching needs real intelligence, not artificial intelligence’, Philosophy of Coaching: An International Journal, 9(2), pp. 6–15. Available at: https://philosophyofcoaching.org/v9i2/02.pdf.

Bhardwaj, C. (2024) ‘AI in Sports – how is artificial intelligence redefining the sports industry? Real-world examples’, Appinventiv. Available at: https://appinventiv.com/blog/ai-in-sports/.

Bridgeman, J. and Giraldez-Hayes, A. (2024) ‘Using artificial intelligence-enhanced video feedback for reflective practice in coach development: benefits and potential drawbacks’, Coaching: An International Journal of Theory, Research and Practice, 17(1), pp. 32–49. 

Couzens, A. (n.d.) ‘Coaching in 2030: how artificial intelligence will change our profession’, SimpliFaster. Available at: https://simplifaster.com/articles/artificial-intelligence-developments-coaching/.

FasterCapital (2025) ‘Sport coaching AI: game changing: how sport coaching AI is revolutionizing team performance;, FasterCapital. Available at: https://fastercapital.com/content/Sport-Coaching-AI--Game-Changing--How-Sport-Coaching-AI-is-Revolutionizing-Team-Performance.html.

Jaswal, R.S., Stergiou, P.  and Katz, L. (2022) ‘How Canadian high-performance coaches adopt and implement technology: exploring the antecedents of intra- and interorganizational trust, technological proficiency, and subjective norms and social influence on technology adoption’, International Sport Coaching Journal, 9(2).

Mazurova, E., Standaert, W., Penttinen, E. and Ter Chian Tan, F. (2021) ‘Paradoxical tensions related to AI-powered evaluation systems in competitive sports’, Inf Syst Front, 24, pp. 897–922. Available at: https://link.springer.com/article/10.1007/s10796-021-10215-8.

Sperlich, B., Düking, P., Leppich, R. and Holmberg, H-C. (2023) ‘Strengths, weaknesses, opportunities, and threats associated with the application of artificial intelligence in connection with sport research, coaching, and optimization of athletic performance: a brief SWOT analysis’, Frontiers in Sports and Active Living, 5. Available at: https://www.frontiersin.org/journals/sports-and-active-living/articles/10.3389/fspor.2023.1258562/full.

Taylor-Price, J. (2024) ‘How AI is changing gymnastic judging’, MIT Technology Review. Available at: https://www.technologyreview.com/2024/01/16/1086498/ai-gymnastics-judging-jss-world-championships-antwerp-paris-olympics/.

Woltmann, L., Ferger, K., Hartmann, C. and Lehner, W. (2023) ‘JumpXClass: explainable AI for jump classification in trampoline sports’. Available at: https://dl.gi.de/server/api/core/bitstreams/7661a84b-5c65-4b6e-ad93-2fff2fee2231/content.

Explore research with, about and led by children and young people at the Children's Research Centre website.

Like this? Try a free course on coaching

• Exploring sport coaching and psychology

  This free course, Exploring sport coaching and psychology, investigates how scientific and management ideas contribute to success while also taking you on a journey through unique sporting case studies and insights that will change how you view and study sport. You will consider how the mind, the body, the environment and training techniques all...

  Learn more to access more details of Exploring sport coaching and psychology

  

• Supporting female performance in sport and fitness

  Do you support women’s performance in sport or fitness as a coach or parent? Are you a female athlete who wants to optimise your performance using recent research insights? If so, then this free course is for you. It addresses the problem of sports science research traditionally being male dominated and the assumption that if it works for men ...

  Learn more to access more details of Supporting female performance in sport and fitness

  

• Coaching others to coach

  Are you responsible for supporting sport and physical activity coaches to become inspirational and excellent in their practice? This course is designed for coach developers, educators, mentors and tutors who support coaches to improve. You will explore how people learn, what are the skills and qualities of an effective coach of coaches, as well ...

  Learn more to access more details of Coaching others to coach