How to Regulate Artificial Intelligence? A Standardisation-based Approach
The European Union (“EU”) has been engaged in a regulatory process in order to set up a balanced legal framework for artificial intelligence (“AI”), both protective for users and innovation-friendly for designers and providers. The project aims to analyse this ongoing process following a standardisation-based approach.
Under the 2021 EU proposal of “AI Act” which provides for a horizontal regulatory framework for AI systems placed on the single market, a key-role is given to standards, understood as voluntary technical or quality specifications with which products or services may comply. Pursuant to Article 40 of the regulation proposal, AI systems that conform to expected new EU harmonised standards would benefit from a presumption of conformity with the AI Act requirements. It means that instead of complying with EU binding statutory law, it would be sufficient, to a certain extent, for providers of high-risk AI systems to conform with harmonised standards accepted by the EU jurisdiction.
Against this background, a twofold issue of legitimacy and efficiency of the future AI regulation within the EU should be raised. To what extent can (current and future) standards applicable to AI systems – placed and used on the EU market – be sufficient to ensure compliance with and promotion of EU values (e.g. democracy, the rule of law, access to justice) and fundamental rights (e.g. protection of privacy and personal data, non-discrimination, freedom of expression) enshrined in EU law? More broadly, the research should make it possible to determine whether there is a regulatory specificity for AI systems and, if so, to draw consequences for the production and application of standards and (soft/hard) law requirements in the field.
Marion Ho-Dac is professor of private law at the University of Artois (France), on delegation at the Academy of International Affairs NRW.
She holds a PhD at the crossroads of EU internal market law and private international law (La loi du pays d’origine en droit de l’Union européenne, Bruylant, 2012), and was a researcher in comparative law at the Court of Justice of the European Union before joining the academic world.
Her teaching and research focus on the governance of the digital single market, EU civil justice and, more globally, the legal governance of international private relationships in a context of growing influence of new technologies which requires a renewed protection of individuals (e.g. consumers, litigants, citizens, children). Marion Ho-Dac also participates, on behalf of AITI, in several working groups led by standardization bodies in the fields of digital technologies and AI.
She is a member of the Centre Droit Éthique et Procédures (CDEP UR 2471) of Artois University and sits on various scientific and editorial boards (e.g. European Association of Private International Law, AI Transparency Institute, Revue Trimestrielle de Droit Européen).