Senate rapporteur promises to send text in April, but analysis face challenges

02/26/2024

---

Senator Eduardo Gomes — Foto: Edilson Rodrigues/Agência Senado

Pointed out by Senate President Rodrigo Pacheco as a priority for the first half of the year, the bill authored by him that regulates artificial intelligence is expected to advance soon in the Upper House. The rapporteur of the matter, Eduardo Gomes, promises to deliver the report for voting in April. From there, lawmakers’ assessment is that the challenges of the text will only be beginning, and the conclusion of the analysis in Congress may be postponed until after the municipal elections to be held in October.

The regulation was advocated at the beginning of the year not only by Senator Pacheco but also by Chamber of Deputies Speaker Arthur Lira and the president of the Superior Electoral Court, Alexandre de Moraes, who urged Congress to act, claiming there are risks to the elections.

Eduardo Gomes said that it is not possible to improvise in the face of a topic that is still unknown to many. For him, it will be up to the electoral court and the regional courts to regulate the 2024 election based on the current legislation. “Voting before the elections doesn’t make sense. Maybe it makes sense to solve minor clashes, but it won’t solve the important issue of a safe law. What defines the time is the security to do it,” he said during an event last week.

To Valor, Mr. Gomes emphasized that other countries have been cautious about the topic to preserve their innovation capacity. “The text will undergo several changes. We have to define where Brazil will be, whether as a consumer or formulator of AI policies. Speed is needed, but not in a way that makes legislation inefficient. The intention is for everything to be resolved in the second semester, until COP-29 [in November],” he said.

The feeling is shared by other lawmakers. Senator Nelsinho Trad said that it is necessary to respect “the rapporteur’s time” due to the complexity of the issue. “The debate can be exercised and move forward in parallel,” he said.

There are more than 50 bills to regulate AI in Brazil. The initiatives are recent—the oldest one date back to 2019.

Among the main points of the proposals is the use of AI during the electoral period. With the advancement of technology, it is now possible to produce extremely convincing fake audiovisual content, the so-called “deep fake.”

In one of the most recent cases, an audio file created by artificial intelligence falsely portrayed the mayor of Manaus, David Almeida, swearing at teachers to harm him. The case is being investigated by the Federal Police.

As the minority leader in the Lower House, Congresswoman Bia Kicis wants to hold public hearings on the subject parallel to those of the Senate and is studying the possibility of creating a joint parliamentary front on artificial intelligence. The aim is to be prepared when the proposal from the other Upper House arrives.

“It will be a theme that will govern the next decades. We don’t need to talk only about fraud. AI has to go far beyond that, in education, medicine, there’s so much that will be impacted by it. I want to get into this topic so that we [from the opposition] don’t remain oblivious, because otherwise, we’ll be bulldozed,” said Ms. Kicis.

The lawmaker criticized the lack of regulation in Congress, which, in her view, allows the TSE to create the rules.

The Lower House has already passed a bill in 2021 to regulate the use of AI. However, the text has been stalled in the Senate since then. Senator Pacheco’s proposal, currently seen as the one with the greatest chance of advancing, was formulated based on a commission of legal experts in 2022 to consolidate the various initiatives on this matter in Congress.

“Now we have to wait for the bill to be approved [in the Senate]. It’s no use doing another one,” said Ms. Kicis.

For lawyer Renata Schuch, a specialist in regulatory law, the indiscriminate use of AI can shape the outcome of elections. “Congress’s rapid regulation is imperative. The wording brought by the bill is already capable of regulating and punishing certain behaviors that may be seen as contrary to the fairness of the electoral process,” she said.

Gabriel Cosme de Azevedo, a specialist in digital law, emphasizes that although the bill has a one-year validity after publication, its effects “have room to be immediate” in terms of awareness and search for future compliance by AI operators. “The law would expand the possibility of gathering cases and improving provisions involving AIs in the electoral context.”

Ticiano Gadêlha, a lawyer specializing in Intellectual Property Law, said that, as in other countries, AI can be used by the TSE itself to ensure greater accessibility in the electoral process and even combat fake news by identifying and blocking such information. “It is necessary to ensure that the technology is used fairly and safely, respecting the rights and, of course, the privacy of voters.”

For him, “technology, as a rule, progresses more rapidly than the law, which means that regulation has a great risk of being obsolete from the start.” “The versions we call intelligent are updated at a very fast pace, so we must have a non-prohibitive regulation and one that punishes any excesses.”

*Por Julia Lindner, Caetano Tonet — Brasília

Source: Valor International

https://valorinternational.globo.com/