Bhargavi Ganesh (University of Pennsylvania), Jose A. Guridi (Cornell University – Department of Information Science), and Thiago Moraes (Vrije Universiteit Brussel (VUB)) have posted Redesigning AI Regulatory Sandboxes for Generative AI: Beyond Traditional For-Profit Innovation on SSRN. Here is the abstract:
Generative artificial intelligence (AI) is expanding the capacity to develop and deploy systems beyond conventional firms, while governance and compliance capacity remain uneven. Regulatory sandboxes could support these innovators, but their accessibility to non-market actors remains poorly understood. We conduct a comparative mapping of 16 AI-relevant regulatory sandboxes across 13 jurisdictions, using publicly available official documents to examine program goals, eligibility criteria, eligible actor types, and admitted participants. Regulatory learning is a stated objective in all programs, and thirteen include societal benefit among their eligibility criteria. Yet businesses are the largest admitted category in every program with published cohort data, including those open to non-market organizations. No program explicitly extends eligibility to individuals or informal collectives, and none of the published cohorts includes non-institutional individuals. These findings reveal a gap between formal eligibility and realized participation. We discuss resource and capacity constraints, legal requirements and policy choices, and access and outreach as potential explanations. We argue that generative AI makes this gap increasingly consequential by weakening the relationship between organizational form and productive capacity without eliminating differences in governance capacity. Participant composition therefore matters both for access to innovation support and for the scope of regulatory learning. We propose selection criteria centered on project characteristics, support for developing governance capacity during participation, and differentiated access pathways. The paper contributes to innovation policy by examining how instruments designed around conventional organizational signals can overlook emerging innovators and limit the contexts from which regulators learn.
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum
