The purpose of this document is to present a first attempt to define what to consider when establishing an EU regulatory sandbox from a perspective of management. We do this by summarising and synthesizing both the principles laid out in the relevant EU regulations, our own case-based research and by including other sources we found relevant to governing regulatory sandboxes.

In short, we present an intermediate governance model for regulatory sandboxes. By intermediate we mean both as a step between the current situation and the future when best practice for conducting regulatory sandboxes is established, as well as to show shared aspects of governance across the different kinds of regulatory sandboxes. Our primary audience are the authorities that will establish or be involved in the operations of the regulatory sandboxes. The outlined model may of course be adapted by other organisations, as well as to the specific contexts of the case at hand.

This report is a deliverable from the pre-study Forsa (Förstudie om regulatoriska sandlådors roll för AI-baserad innovation, Vinnova dnr. 2024-01668), which is a research project funded by the Swedish innovation agency and conducted by Theo Andersson, Mats Snäll and Jeanna Åkerman at the Swedish Agency for Digital Government and Håkan Burden and Susanne Stenberg at RISE Research Institutes of Sweden. The ambition with Forsa was to explore the role of and cooperation among public authorities in relation to AI-based innovation in regulatory sandboxes.