The Evolution of Technology in Doping Detection in Athletes
At the 2026 Milano Cortina Winter Olympics, Rebecca Passler, a biathlete hailing from Italy, tested positive for a prohibited substance known as letrozole.[1] Passler successfully appealed the result, arguing that the substance arose due to cross-contamination from a spoon used to eat Nutella that was shared with her mother, who took medicine containing letrozole as part of her breast cancer treatment.[2]
For most of anti-doping’s history, a relatively simple question was asked: whether there was a prohibited substance in the athlete’s sample. Answering this question relied on testing an athlete’s biological samples, such as blood or urine testing, to detect doping.[3] However, in recent years, technological innovation and artificial intelligence (AI) have been implemented to more accurately detect doping in athletes.
Specifically, anti-doping methods have evolved to monitor changes in an athlete’s biology over time.[4] One such example is the implementation of an Athlete’s Biological Passport (ABP), which was first approved in 2009. In addition to analysing urine and blood testing, through an ABP, anti-doping authorities such as WADA can monitor selected biological variables over time and look for changes within blood samples that are indicative of doping.[5]
A.R.I.E.T.T.A – the beginning of technology in anti-doping
The earliest integration of artificial intelligence in anti-doping methods was the project of A.R.I.E.T.T.A (Artificial Intelligence Evoking Target Testing in Anti-Doping), which was approved in 2005.[6] The aim of A.R.I.E.T.T.A was to develop a system that showed an athlete’s profile and detected abnormal patterns that were indicative of prohibited substances. Using athletes from the International Biathlon Union, blood data were collected and tracked for haematological and multiple other variables.[7] After entering each athlete’s blood test results into the system, the system would analyse that data over time, identifying patterns and generating profiles for individual athletes or teams. Importantly, rather than looking at a single blood value or test in isolation, the system would calculate a risk score for each sample by weighing multiple markers together, and compare it to how much those markers had shifted from the athlete’s own past results. Using these figures and data, an athlete’s risk score for doping was developed into a risk profile.[8] Then, after uploading an athlete’s results, the system would highlight athletes that produced statistically unusual results, indicating the need for further testing.
How has A.R.I.E.T.T.A influenced the use of technology in today’s anti-doping detection methods?
The A.R.I.E.T.T.A project has formed the basis of the implementation of AI in today’s anti-doping sphere. It illuminated the utility of computerised analysis in examining athletes' blood samples. Today, this system has increased in sophistication and enhanced the athlete biological passport. Instead of merely tracking fixed markers such as haemoglobin and testosterone, it has expanded to detect a variety of chemicals and metabolites, allowing a snapshot of the chemicals present in an athlete’s blood sample.[9] This allows for the detection of a greater range of doping strategies, such as the ability to detect urine swapping and subtle metabolic changes that could mask substance use and that may have bypassed methods from the previous system or human analysis.[10]
In addition to constructing an Athlete’s Biological Passport, implementation of technology such as the Anti-Doping Administration and Management System (ADAMS) has allowed for the tracking of athletes’ locations, travel schedules, laboratory results and Therapeutic Use Exemptions (TUEs).[11] While ADAMS functions as the athlete’s data storage, AI is used as an analytics tool to utilise the data stored within ADAMS, such as an athlete’s location patterns and operates to detect anomalies.
The increasing sophistication of performance-enhancing methods has warranted the development and implementation of technological innovations that are pivotal in detecting doping and preserving the integrity and fairness across competitive sports. This has manifested in technology that allows for the tracking of an athlete’s biology over the course of their sporting career, storing data and evidence from their daily life and creating a holistic profile of an athlete. These developments demonstrate how technology has been pivotal in strengthening the detection of doping.
However, Rebecca Passler’s experience at the Olympics cautions against blind reliance on technological analysis. While her sample contained letrozole, this chemical anomaly had not arisen due to doping but rather through accidental cross-contamination. Thus, while technology has created a more sophisticated and accurate means of detecting foul play, correct interpretation of anomalies remains crucial in maintaining the integrity and fairness of the sporting world.
References
[2] Ibid.
[3] https://www.wada-ama.org/en/athlete-biological-passport
[4] Ibid.
[5] Ibid, and comparing it to the extent to which those markers have shifted from the athlete’s.
[7] Ibid.
[8] Ibid.
[9] AbuHaweeleh, M. N., Hamdan, A., Al-Essa, J., Aljaal, S., Al Saad, N., Georgakopoulos, C., ... & Elrayess, M. A. (2025). Integrating Advanced Metabolomics and Machine Learning for Anti-Doping in Human Athletes. Metabolites, 15(11), 696.
[10] Streun, G., Steuer, A., Ebert, L., Dobay, A. & Kraemer, T. (2021). Interpretable machine learning model to detect chemically adulterated urine samples analyzed by high resolution mass spectrometry. Clinical Chemistry and Laboratory Medicine (CCLM), 59(8), 1392-1399. https://doi.org/10.1515/cclm-2021-0010