At present times, the disruptive impact of artificial intelligence technologies requires prompt actions by regulators to address the risks that AI systems may pose to society. In this context, policymakers in the E.U. and in the U.S. are responding to the challenges of AI regulations through apparently different approaches.
On the one hand, the E.U. has adopted a comprehensive framework of AI regulation that is primarily based on a risk-based approach, which is reflected in the requirements and obligations provided in the AI Act. Building on this approach, the E.U. has made significant efforts to harmonize the tort regime applicable to AI users under E.U. law, as exemplified by the revision of the Product Liability Directive and by the drafting of the AI Liability Directive.
On the other hand, the U.S. has shifted its original approach to the field of AI regulation, by repealing the risk-based regulations enacted by the former Administration. As a result, the U.S. is now leaning towards a tort-based approach, which is fundamentally based on the role of the U.S. liability regime as an incentive to take precautions against the risks posed by AI systems.
From this perspective, the transatlantic framework of AI regulation underlines a clear distinction between the E.U. and the U.S. regulatory approaches: a (mainly) risk-based approach to AI regulation in the case of the E.U.; a tort-based approach to AI regulation in the case of the U.S. However, a more detailed view on the framework of AI regulation highlights that both E.U. and U.S. regulators do not advocate for a singular regulatory approach but rather seek complementarity between right- and risk-based approach to AI regulation.
Expanding on this perspective, this paper addresses the complementarity between these two approaches to AI regulation in the framework of E.U. and U.S. law, grounding the analysis in their policy rationale and economic implications. By relying on economic considerations on the function of the right- and risk-based regulatory approaches, we suggest that E.U. and U.S. regulators should build a complementary framework of AI regulation that is based on three core recommendations: (i) ensure a meaningful degree of public representation in the framework of risk regulation and AI standardization; (ii) invest in the enforcement structure of the AI regulatory framework; (iii) impose a strict liability regime upon AI users.
By implementing these three recommendations, E.U. and U.S. regulators would develop a regulatory framework that respects the economic principles of risk regulation, while promoting a certain, uniform, and socially optimal approach to AI regulation.