As Brazil debates AI bill, calls for facial recognition bans emerge

A group of rights organizations is calling on the Brazilian government to pass the country’s proposed AI regulation and introduce a ban on facial recognition use by law enforcement.

Brazil is currently debating amendments to Bill No. 2338/2023 on the use of AI. Drawing inspiration from the EU AI Act, the rulebook categorizes AI systems according to different levels of risk and aims to establish compliance requirements for AI system providers.

In an open letter, Rights in Network Coalition (Coalizão Direitos na Rede, CDR) says that the bill is a good guide for AI regulation, although the legislation has room for improvement. One of these improvements would be a ban on facial recognition use for public security and criminal justice over its impact on freedom of expression and assembly.

The technology has previously been used to identify protestors who stormed federal buildings in the capital of Brasilia in January 2023, following the defeat of then-president Jair Bolsonaro.

Another reason for the ban is ongoing concerns that facial recognition systems could aggravate existing racial discrimination and lead to errors, the group highlights. The latest example is a man detained by the police at a stadium in Sergipe in April after a facial recognition misidentification. Other cases have also been recorded throughout the country.

“To ensure AI systems promote innovation based on human rights, ethics, and responsibility, it is crucial to establish minimum rules,” says CDR, adding that this does not impede development and innovation. “AI regulation based on rights and attuned to its risks would help to position Brazil as a pioneer in providing and adopting responsible technologies.”

The open letter also accuses technology companies and the private sector of lobbying to prevent the vote on the bill by introducing last-minute amendments and organizing international trips and private events for the senators involved in the debate. The lobbying efforts have postponed the vote on the bills and a new round of public hearings, according to the group.

“In these last hearings, representatives of the private sector insisted on the (fallacious) argument that AI regulation in Brazil would impede innovation in the country, being a huge cost for startups and small and medium-sized companies,” the letter says.

CDR represents a network of more than 50 academic and civil society organizations involved in digital rights, including Access Now.

Brazil’s National Data Protection Authority (Autoridade Nacional de Proteção de Dados, ANPD) published a proposal to amend the country’s AI bill in May. The latest version of the legislation was issued on July 4th and includes basic rights guarantees for individuals affected by AI, defines unacceptable uses of the technology and introduces governance guidance, algorithmic impact assessments as well as measures to combat discrimination.

The bill also introduces civil liability for operators, suppliers and distributors of AI systems, meaning that entities that violate the law and cause harm may be ordered by a court to pay compensation.

The legislation still has a long road ahead with votes pending from both the Federal Senate and the House of Representatives, before being approved by the president. Meanwhile, the ANPD  published a fresh report in June on biometrics and facial recognition.

Article Topics

AI  |  ANPD  |  biometric identifiers  |  biometrics  |  Brazil  |  facial recognition  |  legislation  |  police  |  regulation