Civil society warns of regressions in the draft bill on Artificial Intelligence

Organizations highlight threats to rights protection, concerning erosions, and needs to adjustments to draft bill PL 2338/2023, pending at the Senate

The logos of Google Gemini, ChatGPT, Microsoft Copilot, Claude by Anthropic, Perplexity, and Bing apps are displayed on the screen of a smartphone in Reno, United States, on November 21, 2024. (Photo by Jaque Silva/NurPhoto) (Photo by Jaque Silva / NurPhoto / NurPhoto via AFP)

Civil society organizations are warning of regressions in Draft Bill 2338/2023, meant to regulate artificial intelligence (AI) in Brazil. Presented on November 28th, the report by senator Eduardo Gomes (PL-TO) brings about changes that could weaken legislation, limiting its scope to a few systems and eroding essential governance and rights protection mechanisms.

The first points of concern include the exclusion of systems from the law’s obligations, given the vague wording of Article 1, and the erosion of mandatory risk assessments (Article 12). Labor rights and social participation are also diluted with the removal of guarantees such as human supervision over automated decisions and greater transparency in regulatory cases.

Another point of concern is the pressure to exclude content curation and recommendation systems from the high-risk classification, reducing transparency in digital platforms. The organizations also report attempts to loosen mechanisms against child abuse material and weaken the protection of copyrighted content used in AI training.

These entities ask for the preservation of guarantees and the remedying of gaps so that the law may promote responsible innovation and rights protection. The signatory parties include Conectas, Coalizão Direitos na Rede and Data Privacy Brasil.

Read the full note here.