Regulatory sandboxes for AI: Current situation, challenges and opportunities

Innovators need up-to-date and predictable regulation to develop and deploy their solutions. They may find it hard to understand existing requirements and to establish whether and how these apply to their specific case. Regulators need to guarantee that regulation simultaneously encourages innovation, steers it towards desirable policy objectives and delivers public and market value. … reconciling the needs of innovators and regulators requires a degree of flexibility or space to experiment …’(EC Staff working document Regulatory learning in EU 2023). The AI act envisages one of those novel approaches and provides for the set-up of coordinated AI ‘regulatory sandboxes’ to foster innovation in AI across the EU while guaranteeing appropriate safeguards.

The panel is envisioned to be a continuation of last year’s discussion on the topic at EBDVF 2022 which was also further explored during stakeholder workshops. The aim is to present the development of the regulatory sandboxes for AI in the last year by focusing on their local, as well as transnational dimension in the AI Act and the added value of sectorial regulatory sandboxes. Our expert panelists will leverage their experience, present results, identify challenges, share best practices, propose further exchange and collaboration on the topic.

Key points for discussion will entail:

• The notion of the regulatory sandbox needs to be further crystalised and separated from other experimentation tools due to its regulatory and compliance dimension. The session will attempt to go deeper into substance by focusing on practical questions e.g. on
  • testing parameters (testing in ‘real world conditions’ versus a ‘controlled environment’)
  • incentives for businesses to participate from the viewpoint of compliance presumption, liability limitation, increased accessibility to data, funding,…
• States and sectors are on different levels regarding design and implementation of regulatory sandboxes in general and especially in the field of AI. What can be the role of pilot projects, regulators, other initiatives in the ecosystem like data spaces and testing and experimentation facilities (TEFs),…?
• There is a need for knowledge sharing, talent and capacity building in a number of states in order to successfully utilize regulatory sandboxes for AI in the scope envisioned by the AI Act.

Panelists will kick-off the session with presentations,  followed by an interactive part with Q/A from the session facilitators and the audience. We hope to welcome stakeholders and interested actors from different background and contribute to identifying and overcoming challenges for the establishment and the following successful scaling up of regulatory sandboxes for AI.