This article argues that regulatory sandboxes have become a necessary legislative tool to foster new business models. In the past, some new ideas failed because of the lack of such a legislative tool, shown by the Uber case. AI may play a dual role as an object of regulation and a tool of supervision, including RIA, which poses an additional threat to society at a time when it is becoming apparent that it may escape human control. This requires a reversal of the goals of a sandbox where providing security becomes more important than nurturing novelty. While the EU Commission and scholars encourage trust in AI, distrust should be the guiding principle. In addition, the rule of law may be compromised due to the use of regu-latory sandboxes as a back door for the legislative and administrative authorities to go around the established principles.