 The use of machine learning algorithms has been shown to be effective in detecting and predicting epileptic seizures. However, the accuracy of these algorithms may be affected by the use of unreliable validation procedures. This paper examined the effectiveness of various validation procedures on two datasets and found that the use of more stringent validation procedures resulted in lower accuracy rates. These findings highlight the importance of defining reliable validation procedures prior to deploying any AI-based system in a clinical setting. This article was authored by Sina Shafiazada, John Marco Duma, Giovanni Mento, and others.