Brazilian ANPD launches 2nd edition of its Technology Radar, this time on the subject of biometrics and facial recognition

Brazilian ANPD launches 2nd edition of its Technology Radar, this time on the subject of biometrics and facial recognition

On 24 June, the Brazilian National Data Protection Authority (ANPD) released the second volume of its “Technology Radar” series. The research was conducted by the ANPD’s General Coordination of Technology and Research (CGTP) and, this time, addresses biometric and facial recognition technologies, also bringing up issues related to the applications and impacts of these technologies in various sectors, as well as discussing the impact of the use of artificial intelligence (AI) and the risks and challenges associated with the privacy and protection of personal data.

The study begins by conceptualizing biometrics as the technical analysis of the physiological and behavioral characteristics of individuals using mathematical and statistical methods. It explains that these characteristics can include fingerprints, facial recognition, iris, hand geometry and voice, among others. In this sense, the accuracy and reliability of biometric analysis increases as the amount of data present in the biometric sample grows, providing greater quality for the analysis. Here, the ANPD justifies the popularization of biometrics based on its ability to increase security, whether in electoral and financial transactions, border and airport control, or in access to equipment and systems.

Next, the ANPD looks in particular at the issue of AI when used to support the innovations arising from facial recognition. Here, the study brings up machine learning and deep learning as tools capable of optimizing traditional facial recognition systems. At this point, however, it emphasizes the possible risks of invasion of the privacy and civil rights and freedoms of data subjects.

Elsewhere, study discusses the cultural and social biases and norms of the individuals responsible for processing biometric data, which can be reflected in algorithms and learning models, leading to discriminatory effects of a racial, social, ethnic, gender or economic nature. Also, in the chapter “Public security and migration control”, the document considers aspects related to the ability of facial recognition systems to correctly identify a person, which can vary according to the method and technology used by the developer, or the diversity of the population to which they have been applied. He gives an example of this with a case that occurred in Rio de Janeiro, where a woman was mistakenly identified and arrested on the basis of an erroneous biometric identification.

He also addresses the important debate about the introduction of facial recognition in the school environment. On this topic, it raises the use of biometrics to automatically record student attendance, data which, in turn, could be used to support policies to optimize the management of the school environment, combat truancy and ensure safety. However, it warns against processing sensitive data, given the greater vulnerability of data subjects, most of whom are children and adolescents, under the terms of the LGPD.

Finally, it highlights the work of technology companies that have plans to adopt passwordless authentication, using biometrics instead. He points out that although Brazil does not yet have specific federal legislation on video surveillance and facial recognition, there are legislative proposals in the National Congress. Article 4 of the LGPD establishes rules that must be observed for the processing of sensitive data, even for public security, including due process of law, the general principles of protection and the rights of the data subject.

The document can be accessed via the link: Radar Tecnológico nº2 ANPD

Note: For quick release, this English version is provided by automated translation without human review.