Smarter AI innovation happens in a box

COMMENTARY By Thiago Guimaraes Moraes

Safe and responsible AI development requires careful testing and regulation. That’s why AI regulatory sandboxes—controlled environments for experimenting with innovative technologies and data practices—are a promising tool for the Global Majority to allow for technology leapfrogging while fostering responsible innovation and ensuring that new systems respect fundamental rights.

These sandboxes serve a dual purpose: They are secure environments for testing technologies, and they are collaborative frameworks that allow regulators, innovators and other stakeholders to assess technologies against existing laws and ethical standards. Sandboxes are time-limited, iterative processes designed to evaluate innovation while ensuring regulatory alignment.

The benefits of sandboxes extend beyond innovation. They provide a structured mechanism to balance the trade-offs between fostering technological advancement and maintaining regulatory compliance. By encouraging the pooling of datasets and expertise, sandboxes allow participants to explore new possibilities while addressing concerns around privacy, fairness and accountability.

For the Global Majority, regulatory sandboxes offer a strategic approach to bridge the gap between innovation and compliance. By fostering localized benefits and demonstrating the tangible value of AI investments, sandboxes can attract international partnerships and funding. Regional pilots, enabled by these environments, showcase the potential of AI to address local challenges, from optimizing public service delivery and access to advancing health-care solutions tailored to specific contexts.

For example, in Africa, the nonprofit Datasphere Initiative has organized sandbox discussions to build a pan-African community to enable innovative cross-border data governance solutions. The first iterations focused on operational sandboxes, and as the initiative moved forward, it facilitated discussions to create regulatory capacity and connections via roundtables.

Implementing sandboxes in the Global Majority is not without challenges. Limited resources often hinder the establishment of effective initiatives. Without proper planning and capacity-building, these efforts risk becoming time-consuming exercises with little regulatory learning or meaningful outcomes. Another challenge is ensuring that sandboxes are not just experiments but effective platforms for real innovation. Poorly designed sandboxes can lead to outcomes that neither advance regulatory understanding nor deliver societal benefits. To address these challenges, the United Nations Economic Commission for Latin America and the Caribbean developed a tool to assess a country’s readiness to implement a sandbox.

To maximize the benefits of regulatory sandboxes, the Global Majority must adopt a strategic approach that leverages partnerships and builds on existing strengths. Consortiums involving governments, private sector players and international organizations can pool resources and expertise, addressing resource constraints of developing economies. In an increasingly interconnected world, the Global Majority has a unique opportunity to lead by example, demonstrating how AI can be harnessed to address local challenges while adhering to global principles. 

Thiago Guimaraes Moraes is a joint-degree Ph.D. Candidate in law at University of Brasilia and Vrije Universiteit Brussels and a PhD fellow of the Digital Governance cluster at the United Nations University’s Institute on Comparative Regional Integrated Studies.