“The state and regulation cannot just be reactive. They must be part of the innovation process. Smart innovation needs smart regulation.” On the strength of this observation, the “connector” of the National Commission for Data Protection (CNPD), commissioner Alain Herrmann, announced on Tuesday evening--at the end of the inaugural edition of the Fedil AI Excellence Awards on the Nexus Luxembourg main stage--the creation of a Luxembourg AI sandbox.
Named “Regulation Meets Innovation” (Remi) and in collaboration with the AI Factory led by 
This is only the second attempt after the one announced a year ago. This one seems to be more collaborative and educational, with the expected involvement of the University of Luxembourg.
How will it work? An entrepreneur who knows that their technology will face compliance challenges can receive support to ensure that they are in line with the forthcoming rules, before their solution is put on the market. “The sandbox does not provide certification. It will provide an assurance of compliance… Notified bodies will provide the stamp,” explains Herrmann. This approach will make it possible to test specific elements over short periods, in order to generate useful lessons for the market as a whole. “The idea is also not to make something last for three or four years, but to be short and fast. And the lessons learned will be shared in general guidelines.”
The CNPD is taking on the central role that the regulator is going to give it in this process: it becomes the single point of contact, and will be responsible for the national coordination of AI regulators, as provided for in the European bill currently under discussion. If all goes well, sector-specific regulators (for financial or healthcare products, for example) will complete the set-up, in a multi-level approach.
This article was originally published in French.
