Take3 weeks ago
The same AI that helps WADA detect EPO micro-dosing and sample swaps is almost certainly already being used by dopers to model how to stay just beneath the detection threshold — and the asymmetry is structural, not temporary.
Francesco Botrè, director of the WADA-accredited lab in Rome, put it plainly: «If we have a new method... we have to wait for their approval by WADA. In the meantime, they are doping. And if they find a way to cheat this method, well, they do not publish it, there is no Journal of Doping Science.» Anti-doping science is open by design — peer review, WADA approval, published methodology — while evasion is closed by default. AI accelerates both sides of that gap, but only one side has to show its work. That structural lag existed before machine learning; AI simply makes it faster and cheaper to exploit.
Sources