Skip to main content Skip to main navigation



AI-based detection of steroid sample swapping in the anti-doping area and analysis of the associated decision-making processes

AI-based detection of steroid sample swapping in the anti-doping area and analysis of the associated decision-making processes

  • Duration:

Anti-doping analyses are important in maintaining fairness and integrity in sports, with robust measures required to fight against cheating and doping. Recent advancements have shown the potential for new approaches, such as machine learning, to play a pivotal role in these efforts. One significant challenge in anti-doping is the issue of sample swapping, where athletes exchange their urine samples with others to evade detection. To address this, our project aims to leverage machine learning algorithms to enhance the sensitivity and accuracy of finding sample swapping by improving the targeting of DNA identification techniques, which confirms the identity of the sample's donor, thus detecting any exchange of urine samples between athletes. The project integrates the previously developed algorithm from the 3S-project that flags potential sample-swapping cases into existing anti-doping protocols, to be used both retrospectively and prospectively. 3S-II aims to develop and integrate a user-friendly tool that provides anti-doping authorities with a robust mechanism for monitoring and assessing athletes' samples. The tool will generate a similarity score between a given sample and other samples within an anonymized longitudinal profile of an athlete, aiding in detecting inconsistencies indicative of doping. Decision-making processes in anti-doping are complex and decisions regarding the interpretation of doping samples need to be comprehensible and precise. While AI poses tremendous possibilities to improve the correctness of sample swapping detections, it will also change the decision-making process. Within 3S-II, we will study how the implementation of the AI-based tool to detect sample swapping will influence the decision-making process. The project will follow a design science research methodology, encompassing several stages, including exploration of the current passport assessment processes (decision-making), design and development of the algorithm, its demonstration through standalone software, and evaluation of ist impact on decision-making processes in anti-doping. This comprehensive approach aims to significantly improve the ability of anti-doping organizations to identify and prevent sample swapping, ultimately leading to cleaner sports and fair competition.


World Anti-Doping Agency (WADA)