As in many other areas, artificial intelligence (AI) has taken over sport in an incredible way. In just a few years, all aspects of sport, both professional and amateur, have been affected by AI in one way or another. So much so that today’s athletes, referees, coaches and fans have a totally different relationship to sporting activities compared to a few years ago*. The same goes for medical staff, nutritionists and psychologists. Artificial intelligence has most certainly revolutionised competitions, sports medicine, training, refereeing, sport psychology, nutrition, as well as how athletes, fans and society relate to one another. Not to mention the impact of AI on sports betting and the huge business behind it. There are positive aspects to all this, particularly for effective refereeing decisions or detecting injuries or harassment of athletes. But there are also negative aspects that can seriously compromise the wellbeing and safety of athletes as well as the integrity of the competitions.
Artificial intelligence in professional sports
These days, the task of identifiying and selecting sporting talent has changed. No longer based on observing sport competitions, the task now involves algorithms providing the necessary information about each athlete, their chance of success in one competition or another, and the profit that can be made from their participation and sale later on. This is great news for clubs and sports representatives. Not so much for athletes, as we will see later.
Of course, the data the algorithms work with comes from the athletes themselves, whose movements, endurance and skills are analysed in and out of competitions. However, athletes do not possess this data, so it can be used without their consent and even against their interests. Consider, for example, if biological data, collected through different AI systems, indicates decline in an athlete’s performance or predicts that (s)he will stop performing at the same intensity at a certain point in time. This data can be used by the sports club to renegotiate or even terminate the athlete’s contract. In other words, the athlete’s data, initially collected to improve that athlete’s performance or the tactics and strategies to be used in competitions, ends up being used against the interests of its owner, who has given it away free of charge. Can athletes refuse to have their data analysed, or would they then be accused of going against the interests of their club, their team, even their sponsors?
Something similar is happening with refereeing. In almost all sports, referee support systems have been implemented with different technologies, and they greatly facilitate the work of referees, judges, and officials. This means that when AI is applied, human refereeing is reduced to approving decisions recommended by the technology, which happens a lot in tennis and artistic gymnastics. Is the future of refereeing in doubt? Will referees be replaced completely by AI systems, as FIFA has already announced?
The continuous advance of AI in sport brings to the fore many theoretical and practical questions about the role AI plays in sport.
Sport and humanism
It is important to consider that the exponential growth of AI in sport seriously jeopardises one of the fundamental postulates of our conceptualisation of sport until now — that sport is essentially a competitive activity between human beings. In that sense, it is not pointless to ask whether increasingly incorporating this technology into sport might not in fact mean that sport becomes denaturalised, and that we increasingly have a competition between man and machine. Of course, until not so long ago, referees and judges in different sports competitions were considered an additional element in sport. That is to say, those responsible for enforcing the rules were also part of the game, so the possibility of bad decisions happening as a consequence of human fallibility was accepted. It is true that these situations create much inconvenience and generate doubts about the integrity of referees and judges. But it is equally true that dehumanising decisions results in adjustments to techniques and exercises, particularly in rhythmic gymnastics, in synchronised swimming, and as part of the strategies used in team sports. And with all that, we lose the immediacy and naturalness of sporting skills.
Something similar is happening in training and sports medicine. What was, until recently, under the sole discretion of coaches or sports doctors is now utterly mediated by AI. No doubt this can be very useful, but it is evident that there is less and less room for human decisions. As can be imagined, clubs and sponsors are not willing to invest large sums of money in a technology only to have it ignored by those who would initially benefit from its recommendations. But don’t we then dehumanise sport, depriving it of its distinctive feature? Is it possible to conceive of a competition between humans and machines, or will the sport of the future be simply another spectacle?
Following on from that, on an immediate practical level, it is important to ask ourselves what this situation means for the future of coaches, referees, medical staff, nutritionists and other kinds of professionals who will see the scope of their activities reduced, or eliminated due to being replaced by AI.
Protecting athletes and competitions
Now, the increasing use of AI in sports also has another lesser-known aspect which poses more dangers for athletes’ wellbeing and the integrity of competitions. On the one hand, it is well known that AI feeds on a huge amount of data that is collected from the athletes themselves through various so-called “wearables”. Much of this data is extremely sensitive since it includes biometric information and is therefore strictly private. In truth, the data is classified as high-risk data in Article 6.2 of the EU AI Act and, as such, is subject to strict risk management and an exhaustive set of risk mitigation measures.
In addition, the way in which the data is collected, and how it is used — at times against the interests of the athletes themselves — is a sensitive issue. Some of the fundamental rights of not only athletes but also referees and other people involved in sport are under threat.
On the other hand, it is necessary to take into account that along with AI technology’s potential for improving sporting performance will emerge new forms of doping, or technodoping, with consequent risks to the integrity of competitions and athletes’ health. Consider, for example, the increasing use of transcranial direct current stimulation techniques (tDCS) in sports, the effects of which could significantly increase with the incorporation of AI.
Similarly, the increasing use of AI in sports also creates other risks related to the integrity and fairness of sports competitions such as the clearly unequal distribution of not only this technology but, above all, constant updates. The resulting risk is the development of first and second class athletes — depending on the level of access they have to the latest versions of AI utilised in their sport.
Some basic recommendations
Several risks associated with the increasing use of AI in sport need to be addressed as soon as possible. And so those responsible for governance in world sport — particularly the International Olympic Committee and the main international federations such as FIFA, World Athletics etc. — need to urgently adopt some fundamental measures to safeguard athletes including, at the very least, the next generation.
(i) Safeguarding minors. Collecting and processing biometric data of underage athletes has to be strictly prohibited.
ii) Privacy. There should be limitations on the data collected and there should be protection clauses to guarantee that the data remains at the disposal of its owner at all times.
iii) Autonomy. The increasing use of AI in sports must be accompanied by a guarantee that athletes, referees and coaches have the autonomy to make decisions in their own field of action without AI recommendations being imposed upon them.
iv) Equality and fair distribution. All athletes must be guaranteed equality of access to AI systems and their updates so that no one can be disadvantaged by missing out on the competitive benefits it can offer.
v) Human control. AI systems must be subject to human control and actions. Their purpose must be the protection and wellbeing of athletes, referees, coaches, and other personnel involved in sport.
Finally, AI currently offers huge opportunities for all aspects of sport, but it also brings significant risks to the wellbeing and rights of athletes. Taking advantage of the opportunities while safeguarding against the risks is the responsibility of the governing bodies of world sport, and they must act as soon as possible.
Image: Photo by Surface on Unsplash
* In the Italian version of this article, the masculine plural is used for brevity. This is not intended to neglect or underestimate the growing importance of women in all fields of sport.
