Groundwater contamination by chemical toxins poses critical threats to water security globally. The Mediterranean basin exemplifies this challenge through widespread quality deterioration driven by agricultural intensification, industrial expansion, and climate variability. This study develops an integrated artificial intelligence (AI) framework to classify World Health Organization (WHO) guideline exceedances across six Mediterranean countries: Morocco, Palestine, Egypt, Tunisia, Italy, and Spain. The research compiled 330 groundwater samples collected between 2018 and 2023, analyzing Arsenic, Cadmium, Chromium, Lead, Nitrate, Fluoride, and trace metals. A three-stage AI workflow employed K-Nearest Neighbors (KNN) clustering, Decision Tree (DT) analysis, and Random Forest (RF) modeling. KNN achieved 78.8% accuracy, separating low-risk countries (Italy, Spain) from high-risk groups (Egypt, Morocco, Palestine, Tunisia). DT attained perfect classification performance (100% classification accuracy, Matthews Correlation Coefficient = 1.00) with significant country differences (χ2(5) = 17.21, p = 0.004). Italy exhibited 6% exceedance rate while Tunisia showed 83%. RF demonstrated superior performance with 98.1% classification accuracy and an area under the curve of 0.99, generating probabilistic risk scores: Tunisia (85%), Egypt (75%), Morocco (62%), Palestine and Spain (50%), and Italy (0%). The framework successfully discriminates contamination patterns across diverse contexts, providing interpretable country-level risk assessments for targeted groundwater management. This methodology advances operational water quality monitoring with direct applicability to water-stressed regions globally.