The Brazilian Data Protection Authority (ANPD) has published the first monitoring report on its Artificial Intelligence Regulatory Sandbox, an initiative that brings together innovative projects in a controlled and supervised experimental environment to assess AI solutions involving the processing of personal data in compliance with the Brazilian General Data Protection Law (LGPD).
The Brazilian Data Protection Authority (ANPD) has published the first monitoring report on its Artificial Intelligence Regulatory Sandbox, an initiative that brings together innovative projects in a controlled and supervised experimental environment to assess AI solutions involving the processing of personal data in compliance with the Brazilian General Data Protection Law (LGPD). The report covers the first of six testing cycles planned for the project, which began on March 18, 2026, and notes that this initial stage was preparatory in nature, focusing on methodological alignment, the organization of monitoring activities, and the definition of the technical and regulatory conditions for the subsequent phases.
The companies participating in the sandbox are:
- Metatext: developer of the Guardion.AI platform for AI security;
- Synapse AI: which is developing Trajetto, an intelligent platform for managing passenger flows in railway transportation systems; and
- Prevvine Tecnologia: creator of STAIDOC, a medical artificial intelligence ecosystem based on large language models.
The ANPD is also supported by the technical team of the Center for Artificial Intelligence and Machine Learning at the University of São Paulo (CIAAM/USP), which assists in implementing the project.
During this initial stage, the information collected and the interactions between the participating teams enabled the ANPD to identify key areas requiring further attention, including transparency, data governance, information security, risk documentation, and data protection measures adopted by the participating solutions. The report also highlights opportunities for improvement ahead of the second testing cycle, such as strengthening communication between participating companies and the CIAAM/USP team, improving document-sharing procedures, and developing a protocol for the use of synthetic data within the testing environment.
Sources:
- ANPD Publishes First Results of the Artificial Intelligence Regulatory Sandbox (PT-BR)
- Selected Participants – Artificial Intelligence Regulatory Sandbox (PT-BR)
