Szabályozási kísérletezés és közhatalmi innováció: a regulatory sandbox modell közigazgatási tanulságai
Regulatory Experimentation and Public Authority Innovation: Administrative-Law Lessons of the Regulatory Sandbox Model

Summary/Abstract: This article examines the institution of the regulatory sandbox in the context of the exercise of public authority and the renewal of administrative law. It starts from the premise that the pace of innovation – particularly in the fields of digital technologies and artificial intelligence – consistently outstrips traditional legislative and administrative cycles, creating regulatory uncertainty and new challenges for the protection of the public interest. Against this background, the article argues that the regulatory sandbox should not be understood as a form of deregulation, but rather as a safeguarded, procedure-based experimental framework designed to support regulatory learning and evidence-based rule-making. The study first situates the sandbox within a public law framework, focusing on the role of administrative decisions, supervisory and inspection mechanisms, and judicial review. It then provides an empirical overview of regulatory sandboxes across several sectors, including financial services, energy, healthcare and artificial intelligence, highlighting both converging patterns and sector-specific differences. Particular attention is paid to the European Union’s Artificial Intelligence Act, which introduces a mandatory obligation for Member States to establish national AI regulatory sandboxes, and to the Hungarian implementation of this requirement. The analysis emphasizes that from 2 August 2026 the sandbox will become a compulsory institutional element of AI governance in Hungary. The article concludes that regulatory sandboxes can function as key instruments of a “learning state”, provided that they are embedded in a transparent institutional framework, apply proportionate safeguards, and ensure the public dissemination of regulatory insights beyond the participating entities.

• Details
• Contents