The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

Regulatory Sandbox


-
Added by:   National contact point
Added on:   29 Apr 2026
Updated by:   OECD analyst
Updated on:   17 May 2026

The Regulatory Sandbox is an initiative by Brazil's National Data Protection Agency (ANPD), authorised under Complementary Law No. 182 of 2021. It creates a controlled environment for collaborative experimentation between the ANPD, regulated entities, and stakeholders to test AI and machine learning technologies within a regulatory framework, with a current focus on algorithmic transparency and compliance with Brazil's General Data Protection Law (LGPD).

Name in original language

Sandbox Regulatório

Initiative overview

The initiative is aligned with the Regulatory Sandbox Reference Guide of the Attorney General's Office (AGU) and is authorised under Complementary Law No. 182 of 2021. The current pilot operates under Edital de Convocação nº 02/2025, with selected participants now moving to a training phase covering regulation, AI, and experimental regulatory environments.

The sandbox targets a broad range of stakeholders. Innovative companies and public sector entities can test new technologies whilst mitigating regulatory risks. Researchers can assess AI in real-world scenarios, generating empirical knowledge for scientific and regulatory advancement. Civil society organisations can contribute critical perspectives to ensure collective rights are prioritised. The ANPD itself uses the sandbox to identify regulatory gaps and test approaches to balance innovation with the protection of personal data.

The chosen focus for the current pilot is algorithmic transparency, specifically, identifying technical and administrative measures to comply with Article 20, paragraph 1, and the transparency principle under Article 6 of the LGPD. This addresses key challenges associated with machine learning-based AI, including lack of transparency in automated decisions, discriminatory biases, and misuse of personal data. At the conclusion of the process, the sandbox is expected to inform a regulatory framework better adapted to the opportunities and challenges posed by emerging technologies, ensuring innovation is conducted ethically and in line with the rights of data subjects.