Policy Papers

Takeaways from the São Paulo Seminar on Artificial Intelligence Governance | Suggestions for Brazil and Latin America after the G20 Presidency

  • tech
  • 17 april 2026

The takeaways stem from discussions held during the São Paulo Seminar on Artificial Intelligence Governance, reflecting a dialogue among the Brazilian Observatory of Artificial Intelligence (OBIA/NIC.br), the Center for Artificial Intelligence and Machine Learning at the University of São Paulo (CIAAM/USP), and the International Research Centre in Artificial Intelligence under the auspices of UNESCO (IRCAI). The seminar was convened by the Brazilian Center for International Relations (CEBRI), which acted as a facilitator of the discussions.

The document brings together perspectives from government, academia, civil society, and the private sector to propose a coordinated and forward-looking governance framework for artificial intelligence in the region. The views expressed herein are those of the authors and do not necessarily reflect the institutional position of CEBRI.

Authors: Joao Pita Costa (IRCAI), Mihajela Crnko (IRCAI), Monika Kropej (IRCAI), Joao Paulo Veiga (CIAAM/USP), Cristina Godoy (CIAAM/USP), Luiz Costa (OBIA/CETIC–NIC.BR), Alexandre Barbosa (OBIA/CETIC–NIC.BR), Masa Kovic-Dine (Universidade de Ljubljana), Vasilka Sancin (University of Ljubljana), Laura Escudeiro (CEBRI) and Ana Paula Podcameni (CEBRI).

Main recommendations

The document outlines  a set of guiding principles for AI governance:

  • Human Rights and Inclusion
    AI systems must promote equity and avoid discrimination.
  • Transparency and Explainability
    Systems should be understandable, auditable, and accountable.
  • Accountability Across the Value Chain
    Responsibility must be shared among all actors involved.
  • Risk-Based Regulation
    Governance should be proportional to the level of risk.
  • Adaptive and Participatory Governance
    Policies must evolve with technology and include multiple stakeholders.
  • Scientific Integrity and Data Quality
    Reliable data and rigorous methods are essential.
  • Environmental and Social Sustainability
    AI development must consider environmental and societal impacts.
  • International Cooperation
    Global coordination is key to effective governance.

The takeaways stem from discussions held during the São Paulo Seminar on Artificial Intelligence Governance, reflecting a dialogue among the Brazilian Observatory of Artificial Intelligence (OBIA/NIC.br), the Center for Artificial Intelligence and Machine Learning at the University of São Paulo (CIAAM/USP), and the International Research Centre in Artificial Intelligence under the auspices of UNESCO (IRCAI). The seminar was convened by the Brazilian Center for International Relations (CEBRI), which acted as a facilitator of the discussions.

The document brings together perspectives from government, academia, civil society, and the private sector to propose a coordinated and forward-looking governance framework for artificial intelligence in the region. The views expressed herein are those of the authors and do not necessarily reflect the institutional position of CEBRI.

Authors: Joao Pita Costa (IRCAI), Mihajela Crnko (IRCAI), Monika Kropej (IRCAI), Joao Paulo Veiga (CIAAM/USP), Cristina Godoy (CIAAM/USP), Luiz Costa (OBIA/CETIC–NIC.BR), Alexandre Barbosa (OBIA/CETIC–NIC.BR), Masa Kovic-Dine (Universidade de Ljubljana), Vasilka Sancin (University of Ljubljana), Laura Escudeiro (CEBRI) and Ana Paula Podcameni (CEBRI).

Main recommendations

The document outlines  a set of guiding principles for AI governance:

  • Human Rights and Inclusion
    AI systems must promote equity and avoid discrimination.
  • Transparency and Explainability
    Systems should be understandable, auditable, and accountable.
  • Accountability Across the Value Chain
    Responsibility must be shared among all actors involved.
  • Risk-Based Regulation
    Governance should be proportional to the level of risk.
  • Adaptive and Participatory Governance
    Policies must evolve with technology and include multiple stakeholders.
  • Scientific Integrity and Data Quality
    Reliable data and rigorous methods are essential.
  • Environmental and Social Sustainability
    AI development must consider environmental and societal impacts.
  • International Cooperation
    Global coordination is key to effective governance.

PUBLICATION EVENT

MORE FROM THE PROGRAM