AI IN CROSS-BORDER PROCUREMENT: LAW AND ETHICS

Abstract

The dynamic digital transformation of public procurement via Artificial Intelligence (AI) creates a critical regulatory disparity with the World Trade Organization Government Procurement Agreement (WTO GPA), which mandates transparency and non-discrimination. This study analyzes the ethical and legal conflicts arising from the cross-border application of high-risk AI systems. The core issue is an ethical and legal trilemma: algorithmic bias, opacity (‘Black Box’), and the need for Explainability (XAI). These generate latent discrimination, violating GPA Article IV and undermining the right to an effective remedy (Article XVIII). A comparative analysis of international regulatory models (EU, UK, US, South Korea) shows convergence on mandatory Human-in-the-Loop (HiL) control and stricter algorithmic accountability. Traditional post-hoc XAI methods are insufficient for legally meaningful justification in high-risk sectors. The paper justifies transitioning to the ‘Justified AI’ (JAI) standard, ensuring inherent interpretability as the sole mechanism for guaranteeing digital accountability and full compliance with due process in cross-border public procurement.