Algorithmic ethics and economic regulation: how do antitrust authorities address the risks of AI?

Abstract

The integration of Artificial Intelligence (AI) into competition authorities to detect anti-competitive practices entails inherent ethical risks, such as algorithmic bias and excessive dependence on technology providers. To mitigate these risks, this article investigates how thirty-five regulatory authorities, ranked in the 2023 GCR Enforcement Rating, address these challenges. A document analysis, carried out using ATLAS.ti25, compares these organizations based on five ethical principles (transparency, accountability, fairness and equity, robustness and security, and privacy), grounded in consolidated frameworks in AI ethics and protection bioethics. The results reveal a critical disparity in regulatory maturity: greater technical rigor is observed in operational ethics principles (robustness, privacy, and security) than in social ethics principles. Substantial deficiencies persist in transparency, accountability, and fairness/equity. This gap points to a deficit in the core principle of explainability (encompassing both intelligibility and accountability). The main cause lies in the absence of clear procedures and limited disclosure regarding the use of AI in the core activities of these organizations. The study concludes that strengthening ethical leadership and establishing organizational accountability mechanisms are essential to ensure the fair and transparent application of AI in economic regulation.