Chasing the algorithm: a law and economics perspective on experimental AI governance

Zarra, Antonella (2026) Chasing the algorithm: a law and economics perspective on experimental AI governance, [Dissertation thesis], Alma Mater Studiorum Università di Bologna. Dottorato di ricerca in European doctorate in law and economics, 35 Ciclo.
Documenti full-text disponibili:
[thumbnail of Combined PDF Bologna.pdf] Documento PDF (English) - Accesso riservato fino a 1 Settembre 2028 - Richiede un lettore di PDF come Xpdf o Adobe Acrobat Reader
Disponibile con Licenza: Salvo eventuali più ampie autorizzazioni dell'autore, la tesi può essere liberamente consultata e può essere effettuato il salvataggio e la stampa di una copia per fini strettamente personali di studio, di ricerca e di insegnamento, con espresso divieto di qualunque utilizzo direttamente o indirettamente commerciale. Ogni altro diritto sul materiale è riservato.
Download (4MB) | Contatta l'autore

Abstract

Artificial intelligence (AI) systems increasingly mediate economic, social, and interpersonal interactions, creating emerging risks that challenge existing legal frameworks. Traditional regulatory instruments—including ex ante regulation, ex post liability, and transparency obligations—were designed for relatively stable technologies and struggle to address the opacity, autonomy, and rapid evolution of AI. This thesis examines how the law can govern risky AI applications under conditions of uncertainty, information asymmetry, and continuous technological change. Adopting a law-and-economics perspective, it evaluates whether conventional regulatory tools can effectively correct AI-related market failures and explores the extent to which experimental governance mechanisms can complement them. The thesis first analyzes the conceptual foundations of AI governance, it then develops a lifecycle-based economic taxonomy of AI-related harms, showing how AI systems generate market failures creating incentives for private actors to underinvest in safety and risk mitigations. Particular attention is given to autonomous vehicles and human–computer interaction. The analysis shows that tort liability faces limitations due to causal opacity and complex multi-actor value chains, which undermine deterrence and complicate responsibility attribution. Similarly, transparency obligations, particularly in the context of generative AI, are constrained by cognitive biases. While disclosure may increase awareness, it often fails to enhance meaningful user agency. To address these limitations, the thesis advances an original framework for understanding experimental governance as an adaptive institutional response to uncertainty. Drawing on original research conducted through policy prototyping and the Open Loop programme, it demonstrates how experimental governance can generate evidence, reduce information asymmetries, and facilitate regulatory learning, provided adequate safeguards against regulatory capture are in place. The thesis concludes that effective AI governance requires a flexible institutional architecture capable of incorporating empirical feedback throughout the regulatory cycle: in that respect, experimental governance can be a means to reduce the expected social costs of regulatory error.

Abstract
Tipologia del documento
Tesi di dottorato
Autore
Zarra, Antonella
Supervisore
Co-supervisore
Dottorato di ricerca
Ciclo
35
Coordinatore
Settore disciplinare
Settore concorsuale
Parole chiave
Artificial Intelligence; Experimental Governance; Law and Economics; EU AI Act; Regulatory Sandboxes; Policy Prototyping; AI Risk Management; Generative AI; Autonomous Vehicles; AI transparency; Experimental Regulation; Market failure; Pacing problem
Data di discussione
9 Ottobre 2026
URI

Altri metadati

Gestione del documento: Visualizza la tesi

^