Author: Stig André Due

Abstract

As artificial intelligence (AI) systems become increasingly integrated into society, the need for effective regulation that balances innovation with public interest is more urgent than ever. The European Union’s AI Act introduces a novel approach to this challenge through the establishment of AI regulatory sandboxes, controlled environments where AI systems can be developed and tested under regulatory supervision. This whitepaper draws on insights from workshops hosted by CAIRNE and FARI in 2024, bringing together diverse stakeholders across the AI ecosystem. It explores the practical design and implementation challenges of AI regulatory sandboxes, with a particular focus on stakeholder engagement, legal frameworks, technical testing, and cross-border harmonization.

Key themes include the importance of taking a use-case-driven approach to sandbox operation, the need for clarity around legal mandates and data protection, and the essential role of trust-building to secure active participation from innovators, regulators, civil society, and experts.

The report also emphasizes the sandbox’s role in generating regulatory knowledge and how that knowledge can be leveraged to support EU-wide harmonization without sacrificing local adaptability. Rather than providing final answers, this whitepaper aims to stimulate a productive dialogue among stakeholders and contribute to the evolving practice of responsible AI governance in Europe.

Read the article