Matheus Puppe speaks to Canaltech about regulating artificial intelligence

Canaltech reported that a commission of jurists formed by the Federal Senate to discuss the regulation of artificial intelligence expressed its opposition to the use of facial recognition systems by the police. Dr. Matheus Puppe, partner in the TMT, Data Protection and Privacy area at Maneira Advogados, was interviewed to comment on the topic.

Senate Committee opposes use of facial recognition by police

By Felipe Demartini | Edited by Claudio Yuge | May 19, 2022

(Image: monkeybusiness/envato)

A commission of jurists formed by the Brazilian Senate to discuss the regulation of artificial intelligence in the country appears unanimous in the idea of banning facial recognition systems by the police. In a review of the debates that took place since April, in Brasília (DF), the idea was contrary to the use of technology, which was assessed as reproducing structural racism.

When presenting the conclusions, rapporteur and jurist Laura Schertel Mendes pointed out the risk of algorithmic discrimination in the use of platforms of this type to recognize suspects and those accused of crimes. The idea is that facial recognition platforms can reproduce the racism embedded in society and do not match the proposal for greater openness and equality that accompanies the modernization of judicial systems.

In addition, she highlighted concerns about the misuse of data collected by such systems, particularly in relation to children and data mining operations. In Mendes' view, there is a need to establish regulations that guarantee safety and are in accordance with the complexity of using such technologies, also taking into account the risks involved in their adoption.

“In addition to the racial biases already mentioned, such [systems] offer intrinsic risks to the freedom and privacy of the general population and to democracy itself”, comments Marcelo Cárgano, lawyer in the area of digital law and data protection at Abe Giovanini Advogados. “Just imagine the indiscriminate use by a government of authoritarian inclinations to persecute its opponents and deploy a mass surveillance platform.”

Senate Commission took a stance against the use of facial recognition by authorities, due to the risk of reproducing structural racism (Image: Geraldo Magela/Agência Senado)

The idea of the commission, which brings together more than 160 experts in 12 panels, is to create subsidies for the production of a replacement bill, which brings together different bills that are being processed in the Senate on the topic. There are three in total, focused on regulation and discipline in the use of artificial intelligence systems, of which facial recognition technologies are also a part.

The proposal follows international concern about the topic, as Cárgano points out. He recalls that other public consultations of this type are taking place, for example, at the European Data Protection Board), which understood that the use of facial recognition has impacts on freedom of movement, assembly and human dignity. “Its use in public spaces has no place in a democratic society, with fears not being hypothetical or exaggerated, but based on concrete examples”, he adds.

Matheus Puppe, partner in the TMT, privacy and data protection area at Maneira Advogados, brings a different vision, in favor of standardizing artificial intelligence programming that can leave prejudiced parameters aside. “We cannot go against the advancement of new technologies, but adopt strict self-regulation measures so that service providers do not feed the system with unfair data”, he states.

The committee formed by the Senate receives external suggestions by email or in person until June 10th. Afterwards, the idea is to help the house in approving a project that guarantees freedom and has clear rules on the use of artificial intelligence and facial recognition systems in the country.

Source: Agência Senado

https://canaltech.com.br/governo/comissao-do-senado-e-contraria-a-uso-de-reconhecimento-facial-pela-policia-216761/