Lack of transparency and national guidelines mark the expansion of facial recognition in brazilian education | InternetLab

In recent years, the debate over technology in education has shifted. In 2023, InternetLab published, with the support of Privacy International, the report “Surveillance Technologies and Education”, which mapped advances in facial recognition in Brazilian public schools and warned of risks to children’s and adolescents’ privacy and non-discrimination. Two years later, the landscape has changed, and public debate on technology in education has moved to generative artificial intelligence tools, which promise to revolutionize teaching and learning. Meanwhile, facial recognition technologies continue to be implemented in schools, largely outside public scrutiny.

It is in this context that InternetLab releases its new report, “Surveillance Technologies and Artificial Intelligence in Education: Promises of modernization, realities of opacity”, published on November 30th. The study broadens the scope to analyze the use of artificial intelligence in public education systems, while maintaining a key distinction: facial recognition remains understood as a surveillance technology, distinct from other AI tools that, when properly implemented, may have pedagogical purposes.

The research shows that behind official narratives centered on efficiency, modernization, and security lies a fragmented landscape marked by the absence of national guidelines. Systems are installed without prior studies or technical assessments, based on unclear legal grounds and with low levels of participation from school communities. In many cases, schools face system failures, overload, and distrust. The narrative of modernization coexists with a limited understanding, among public managers responsible for education policies, of what these tools are and how they function.

The study combined documentary research, freedom of information requests, and interviews with management teams, school staff, and teachers across different regions of the country. It found that the expansion of these technologies has occurred in a disordered manner: artificial intelligence and facial recognition are spreading across public education systems without regulations establishing minimum parameters, amid serious transparency gaps and limited public debate. Some of the main findings help illustrate this scenario:

• Expanding facial recognition, without national standards or evaluation: In 2023, the use of facial recognition was identified in state schools in Tocantins and in municipal schools in the states of Bahia, Ceará, Pernambuco, Goiás, Minas Gerais, Rio de Janeiro, São Paulo, Rio Grande do Sul, and Santa Catarina. By 2025, seven of the eighteen states that responded to freedom of information requests had been implementing facial recognition technologies (FRT) as a state public policy: Alagoas, Amazonas, Goiás, Paraná, Rio de Janeiro, São Paulo, and Tocantins, with Paraná standing out for deploying the system in 1,600 schools. These technologies have been used in relatively similar ways across states, predominantly for student attendance tracking. Alagoas is the only state that explicitly reports additional uses, including access control and student security.
• Other AI tools are already part of everyday school routines: At least three states, Espírito Santo, Rondônia, and São Paulo, have adopted other types of AI technologies in public schools, such as text grading systems and school management platforms, without transparency regarding their purposes or legal bases. In São Paulo, for example, these systems are already operating in more than 5,300 schools.
• Lack of clarity about what constitutes artificial intelligence: Among education managers and professionals, as well as within state governments, there is widespread confusion about what actually qualifies as an AI technology. This lack of clarity is illustrated by states such as Goiás and Alagoas, which, despite confirming the use of facial recognition, denied using artificial intelligence, revealing a lack of understanding of the nature of these tools. This conceptual imprecision undermines the ability to assess risks, define safeguards, or establish mitigation mechanisms, weakening the protection of fundamental rights.
• A gap between discourse and practice: While managers and authorities emphasize gains in efficiency and innovation, accounts from professionals working in schools reveal a daily reality far removed from the promises of innovation and efficiency present in official narratives surrounding facial recognition technologies. Interviews with school teams, teachers, and educators point to technical failures and additional tasks, adding to an already overloaded work environment and the recurrently precarious infrastructure of many schools.
• Lack of transparency: The adoption and implementation of AI and facial recognition technologies are opaque, characterized by generic responses to information requests and a lack of visibility regarding contracts and data flows. The research reveals significant barriers to public transparency, pointing to a scenario of institutional opacity and fragmented responsibilities across federal entities, further reinforced by the failure of nine states to respond to freedom of information requests, in violation of the law.
• Lack of involvement of school communities in decisions about technology adoption: The study also shows that school communities have little or no participation in decisions about adopting these tools. Existing participation spaces within the educational ecosystem are not mobilized to discuss these initiatives or debate privacy and data protection issues. In most schools, these concerns are still absent from everyday practice, revealing a gap between modernization discourse and the lived reality of public education systems. The company responsible for supplying most facial recognition systems did not respond to the research team’s inquiries. This lack of response further reinforces the degree of opacity surrounding the implementation of these technologies.

Given the continued and uncoordinated expansion of these technologies in Brazilian education, the current report reiterates key concerns and recommendations already raised in the previous study. Among the recommendations presented are the development of national guidelines capable of distinguishing pedagogical technologies from surveillance technologies, such as facial recognition, and of establishing minimum parameters for the use of artificial intelligence in schools. These include the preparation of Data Protection Impact Assessments, the adoption of security measures and continuous evaluation processes, and the creation of an interministerial body with civil society participation to monitor the issue.

The study also recommends strengthening the technical and administrative capacities of public management and investing in the continuous training of education professionals, enabling them to critically understand digital tools and their impacts. Finally, it highlights the importance of adequate infrastructure, proactive transparency, and the participation of school communities in technology-related decision-making, which are essential conditions for ensuring that the digital transformation of education advances in a responsible, democratic, and rights-centered manner.

The report is now available for reading.

*This blog post was automatically translated with DeepL and Grammarly, and reviewed by a human editor.