The Bill (PL 2.338/2023) regulating artificial intelligence (AI) in Brazil will be voted on by the Senate’s temporary committee on AI (CTIA), despite criticisms about risks to freedom of expression in the country

The main critics are opponents of President Luiz Inácio Lula da Silva’s (PT) government, particularly targeting articles 14 and 15 of the substitute proposal, which fundamentally alter the initial objective of the regulation.

Article 14 includes the dissemination and even the recommendation of content “on a large scale and significantly automated” in the category of “high risk” AI use, which could open the door for social networks to be classified as “high risk” systems.

Article 15 stipulates that it would be up to a body controlled by the National Data Protection Authority (ANPD) – an autonomous entity under the Ministry of Justice – to judge which systems are “high risk.” Additionally, Article 15 outlines several criteria for the ANPD to make this judgment, including “the risk to information integrity, the democratic process, and political pluralism.”

This last definition, present in item XI of Article 15, has led parliamentarians to accuse Bill 2.338 of attempting to stealthily incorporate the spirit of Bill 2.630/2020 – known as the “Censorship Bill” or “Fake News Bill.”