started · updated
Brazil's CGI.br issues 46 guidelines for social media regulation
The Brazilian Internet Steering Committee (CGI.br) has released a document containing 46 guidelines aimed at regulating social media platforms. The proposal utilizes an asymmetric model, applying different rules based on a company's revenue, reach, and size to ensure proportionality to risk. Key recommendations include requiring foreign platforms to maintain legal representation in Brazil, ensuring algorithmic transparency, and mandating human moderation teams capable of understanding the local cultural context.
Experts have emphasized the urgency of such regulation to combat the risks posed by artificial intelligence. A study by Data Privacy Brasil and Aláfia Lab revealed that nearly 60% of political AI content shared on social networks between January and August 2026 failed to disclose its AI origin. Furthermore, deepfake technology was used in 81% of analyzed cases to manipulate the voices or images of public figures. Researchers argue that increased oversight and media literacy are essential to prevent AI from exacerbating political violence and social biases.
Entities
Aláfia Lab · Blogueiras Negras · Brasil · Brazilian Internet Steering Committee · CGI.br · Congresso Nacional · Data Privacy Brasil · Fernanda Campagnucci · Fernanda Rodrigues · Instituto Sivis · Instituto de Referência em Internet e Sociedade · InternetLab
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 4 SOURCES] Legislation must ensure transparency in social media moderation processes through audits. imprensapublica.com.br · www.jcatibaia.com.br · noticiamarajo.com.br
- [● 3 SOURCES] 38% of Brazilians avoid political discussions with family members at least once a month. imprensapublica.com.br · www.jcatibaia.com.br
- [● 3 SOURCES] Approximately one-third of Brazilians feared reprisals from authorities for criticizing public policies or government agents in the last year. imprensapublica.com.br · www.jcatibaia.com.br
- [● 3 SOURCES] Excessive regulation of digital platforms could worsen existing self-censorship among the population. imprensapublica.com.br · www.jcatibaia.com.br
- [● 3 SOURCES] The internet in Brazil is not a lawless area as criminal legislation and existing regulations like the Marco Civil da Internet apply to social media. imprensapublica.com.br · www.jcatibaia.com.br
- [● 12 SOURCES] AI technology increases political violence related to gender and race via deepfakes and disinformation. ccnnews.com.br · fatopaulista.com.br · itapevinoticias.com.br · mauaagora.com.br · odiariodebarretos.com.br · +5 more
- [○ 1 SOURCE] Deepfake technology was used in 81% of analyzed cases to manipulate the image or voice of public figures. conexaomarilia.com.br
- [● 12 SOURCES] Recommendation algorithms require minimum regulation to prevent discriminatory biases and allow users to opt out of standardized content. ccnnews.com.br · fatopaulista.com.br · itapevinoticias.com.br · mauaagora.com.br · odiariodebarretos.com.br · +5 more
- [○ 1 SOURCE] Nearly 60% of analyzed political AI content on social networks between January and August 2026 failed to disclose the use of AI. conexaomarilia.com.br
- [● 2 SOURCES] The internet in Brazil is subject to existing criminal legislation and regulations like the Marco Civil da Internet. imprensapublica.com.br · www.jcatibaia.com.br
- [● 2 SOURCES] Guidelines include requirements for algorithmic transparency and limits on the use of artificial intelligence. www.brasilemfolhas.com.br · correiodecarajas.com.br
- [○ 1 SOURCE] The document proposes that companies must prove the existence of human moderation teams capable of handling the Brazilian language and cultural context. www.brasilemfolhas.com.br