Ethical AI Regulatory Sandboxes: Insights from cyberspace regulation and Internet governance

This paper explores how theories of cyberspace regulation and Internet governance can provide important lessons for the implementation of ethical-oriented AI regulatory sandboxes, that is, sandboxes that consider ethical aspects in their implementation. The first part provides a brief overview of the challenges in defining the concept of artificial intelligence, and the principles developed to mitigate social impacts arising from the autonomy of AI systems. In the second, the regulatory sandbox is presented as a strategic tool to promote multidisciplinary and multistakeholder engagement. To support this argument, the views of authors who were essential for developing concepts of cyberspace regulation and Internet governance will be looked at, such as Lawrence Lessig, Jeanette Hoffman, Andrew Murray, and Julia Black. Finally, the third part focuses on the Brazilian regulatory environment and its maturity for implementing AI regulatory sandboxes. The article concludes by summarizing the lessons that can be retrieved from the theories analysed for the development of regulatory sandboxes for Artificial Intelligence.