Spot cheaters with the help of artificial intelligence? This is the new objective of the World Anti-Doping Agency (WADA), currently funding several research projects. The objective is to understand whether artificial intelligence can detect signs of the use of doping products that may have escaped human observation.
While the Covid-19 health crisis has caused the cessation of most sports competitions, these will resume in the near future. As the Quebec daily La Presse explains in an article of May 26, 2020, the World Anti-Doping Agency (WADA) took advantage of this judgment to continue its research into the detection of fraudsters. The agency has deployed no less than four research projects integrating artificial intelligence (AI) in Canada and Germany. The aim is to understand whether AI can detect signs suggesting the taking of doping products in cases that have escaped human vigilance.
According to the researchers, the eventual suspension measures for the athletes will not be dictated by artificial intelligence. On the other hand, the latter will designate certain suspect athletes and human controllers will do the checks. In other words, AI will complement human work. It will be a question of combining a lot of data over long enough periods in order to detect what the human brain cannot see.
Analyzing the blood or urine of athletes can detect doping products, but not only. It is also about monitoring many biological markers such as the level of red blood cells or testosterone. However, this type of data is already used by anti-doping agencies through the biological passport. The goal is to allow the AI to track these biomarkers in order to cross-check them with other information. For example, one of the AMA's projects focuses on more accurate detection of EPO. Another project concerns better detection of steroids.
In Montreal (Quebec) a more global project is taking place integrating a larger amount of data. In particular, this concerns athlete location data. However, the researchers are aware that it is incumbent to find a balance between data protection, the protection of individuals and the revelation of the potential of AI, if it exists. We should also mention the fact that none of the projects mentioned integrate AI into the evaluation of the results of athletes in competition. However, this may be the case in the future.
The British daily newspaper The Telegraph already mentioned this WADA-funded research in November 2019. With no less than 300,000 samples tested each year around the world, artificial intelligence could be a real support in the processing of this huge amount of data.
