The Olympic Games are the ultimate sporting event when the ultimate athletes of both men and women of the world come to a single place for the ultimate challenge and battle of the best against the best. These few individuals are rewarded by getting to the level of elite athletes, which involves both genes and hard work, coaches and trainers, as well as the element of luck. Such talent hunt and nurturing of such world-class eggs from the developmental stage is a big problem that the sporting federations and sporting nations face when they dream of success on the world stage. Though talent identification has remained a challenging area for many years, there are signs of technological advancement in the field, particularly artificial intelligence (AI) and machine learning are gradually revolutionizing talent identification.
However, one of the first such implementations can be traced back to the English Institute of Sport (EIS) Elite Coach Development Programme. Currently, since 2019, the EIS has been using the AI athlete assessment tool known as the Performance Genome Project or PGP, which is created by Kinduct Company of Canada. Minimizing the time and cost of an athlete’s preparation, PGP will use physical and cognitive assessments, coupled with the power of big data and computational models, to transform the search for the next Olympic champion.
Before, during, and after activities involving physical demands, adolescent athletes between 12-18 years undergo tests, which involve doing tasks on balance, coordination, strength, speed, and power with the help of electronic devices. Moreover, the activities performed with the tablet enhance the evaluation of the visual processing ability and risk and decision-making. Altogether, PGP’s algorithm used for prediction incorporates more than 3000 parameters for each participant. Using a machine learning algorithm it is possible for the system to compare the athletes with current elite performers of the respective sport and determine those with biometric signatures that are associated with future success trajectories.
Initial trials performed on more than 9000 athletes have revealed quite positive outcomes. In several instances, PGP validation has identified athletes possessing elite capacity, several years before they might have been identified by the traditional coaching talent search. When the high performers are identified at an earlier stage in their development cycle, then fewer resources can be devoted to individuals who are highly likely to reach the pinnacle.
PGP has been implemented within the Team Great Britain performance pathway by UK Sport, and several national governing bodies through Kinduct’s collaboration. The aim is to build a unified system for the development of sports starting from the initial levels of pop-up sports to the podiums of the Olympics. This is a grand plan, and talent identification is but one of the pieces in its scheme. Not only is it important to flag prospects to PGP, but also to monitor their future development stages, as well as train them according to their progress. Here too, AI and data analytics tools try to provide coaches and sports scientists with the informational advantage they need.
Sensors and testing data through athlete management systems such as Kinduct Fusion can be gathered in a way that was not imaginable for even the most knowledgeable coaches twenty years ago. Variables such as the amount of water taken, quality of sleep, training intensity and volume, etc are some of the indicators that will enable the training cycles to be properly tailored and lower the risks of injuries. Such trends and patterns that have not been visible to the naked human eye are identified by machine learning algorithms, hence the belief that future athletes can reach their full potential as has been observed with the available wealth of athlete analytics.
Proposing a technique to predict future champions or even helping shape them, is not a novel concept. For more than a decade, national teams, professional teams, and sporting associations have estimated physical, physiological, and cognitive characteristics in developing athletes. Usually, such testing provided limited and often dated views of what was going on because the testing methods were not rigorous. It was limited by the cost and the fact that information was isolated into organizational silos. Thus, the true innovation of PGP is not in the amount of analytics done but in its applications at the ecosystem level for sporting entities. Thus, making the price lower but at the same time linking the separate strands of the developmental chain, integrated athlete management with the help of AI has long been waiting for its turn to move from the sphere of pure concepts to the sphere of implementation.
Thus, are talent identification systems and performance tracking metrics that are data-centric going to produce a new age of Olympic champions for the country that applies them best? The answer according to British athletics legend Roger Black is yes, the answer lies in the creation of a brand new state-of-the-art centre similar to the famous Australian Institute of Sport which was established in 1981. Regarded as the reason behind the country’s rise to one of the five most dominant Olympic countries after the year 2000, the centralized sports institute concentrated the nation’s best coaching talents, equipment, biomechanics, and medicine to nurture prodigies. Black has the same scenario of a change driven by AI and analytics in the UK as well. Some other authorities suggest that the record medal performance of British cyclists in Rio happened when the team was completely into marginal gains. From additional aerodynamics and washing to increase the number of golds as a result, can Edge, selfokinduct’s AI system reveal more winners like Chris Hoy or Laura Kenny?
To achieve this, there will still be the need for coaches to play the most crucial roles. It is clearly evident that the professional expertise, interpersonal understanding, and on-field behavioral patterns of the head coaches cannot be substituted by any set of algorithms. Biometric monitoring and analysis are intended to act as an extension and aid to human observation and interaction. Instead of eradicating the role of the coaches totally, performance data enables them to make better decisions after analyzing more detailed data. Big data analytics still taps into AI predictions of opportunities and possible threats but there is still the question of intuition acquired through years of practice. The expectation is that advancing AI synergistic with humanistic will bring more athletes to the next level at a faster rate than ever before.
Critics say that the utilization of algorithms undermines the gradual and uncertain process of athletes’ grooming. Some aspects such as dedication, creativity, or even the ability to perform under pressure may be possible to measure but end up not being given much attention. Many worry that a mechanical way will be adopted and implemented. Some critics wonder whether there is any model that can capture the last stages of development curves or the effects of future events as injuries. Some of these are genuine issues, but most of these are easily avoidable or can be managed if deliberate efforts are made in their direction.
This may not be a perfect realization of what advocates of AI-powered talent identification envisioned to be the case once the new technology is embraced. Maybe the machines will find the new Olympic heroes. On the same measure, it may also work to only assist in slightly improving certain components of existing development paradigms. The majority of specialists do not anticipate effects to be either as severe or as beneficial as these two cases suggest. Since there is stiff competition among elite Olympic plans, even tiny advantages that may be secured from new technology form the basis of significant gold medal advantages. If these tools give Team GB that little bit of edge then it could just put them at the top of the medal table when the ultimate selection of athletes gather for Paris 2024.
