Brazil's Senate is set to vote on a bill to regulate artificial intelligence (AI) in late July, but some in the industry say the topic needs more time to be developed and debated.
A senior executive at a major Brazilian telecommunications operator, speaking on condition of anonymity, told BNamericas that he felt it was “premature” to set rules on AI, adding that accountability on the issue remained unresolved and unclear.
The executive said the proposed terms would mean that any entity, company or person that directly or indirectly uses any AI algorithm would be subject to rules on transparency, control and scope, which “doesn't make sense.”
The executive argued that certain laws currently in force in the country, such as the General Data Protection Law (LGPD), the Marco Civil da Internet, as well as the Civil Code and the Consumer Protection Law, already address many of the issues related to the misuse of AI.
On Wednesday, Sen. Eduardo Gomez, the Senate rapporteur for the bill, said the special committee set up to discuss the AI bill would submit its final report by July 17, ahead of a vote in the plenary session of the Senate.
But earlier this week, Gomez said the bill would be voted on in the Senate on June 12. He is also the rapporteur for the bill, which was introduced last May by Sen. Rodrigo Pacheco, president of the Senate and National Assembly.
Media reports have said Lula is interested in introducing the bill at the G20 summit in Rio de Janeiro in November, where Brazil holds the interim presidency.
Discussion
During the June 3 hearing, Senator Gomez cited transparency around AI-generated content and accountability to those who share it as justification for her vote.
“To combat fake news and misinformation, there must also be a verification process – it is essential that whoever puts out content can verify its origin and identity,” he said at the time.
Also attending Monday's hearing was Fernando José García Moreira, coordinator of the AI working group at the Brazilian Television Technology Association (SET).
Moreira called for greater transparency around AI-generated algorithms to prevent content generated by AI from spreading bias and misinformation.
Media, radio and television broadcasters support the bill on ethical and copyright grounds.
“We know that these technologies are trained from large databases. Usually these databases contain copyrighted journalistic works: images, articles, news, texts. The technologies are trained with these materials, without any kind of permission or remuneration. This is a copyright infringement that must be considered carefully,” Andreia Saado, head of Brazil's Radio and Television Broadcasting Corporation (Aberto), said at the hearing.
Meanwhile, tech companies and major telecommunications carriers have urged caution against allowing regulation to stifle innovation.
Conexis, a group representing the largest telecommunications operators, said regulation should be principle-based and flexible, rather than detailed and too specific.
The association also seeks to strike a balance between legal security, protection of the public, and the changing and evolving technological environment.
