Brazil Senate CTIA public hearings on AI Bill PL 2338/2023 — civil society participation (2023)

On 3 May 2023, Brazilian Senate President Rodrigo Pacheco filed PL 2338/2023 — Brazil's framework bill for the regulation of artificial intelligence — a draft that drew directly on a 2021–2022 Commission of Jurists process in which Data Privacy Brasil and allied organisations had contributed substantially to the legal text. For the Brazilian digital-rights civil society field, the bill's introduction opened the central federal legislative engagement of 2023: a multi-month process of public hearings, technical submissions, and coalition mobilisation through the Coalizão Direitos na Rede — the approximately 50-organisation coalition of which Intervozes, Coding Rights, ARTICLE 19 Brasil, and Access Now are members — that constitutes the corpus's first recorded grassroots and civil-society AI-governance engagement event in Brazil.

Context

Brazil had been developing the legal and policy groundwork for an AI framework since at least 2019, when a National AI Strategy was published by the government. The Commission of Jurists appointed by the Senate in 2021–2022 produced a draft bill that was subsequently adapted and filed by Senate President Pacheco in May 2023. PL 2338/2023 proposed a risk-based approach to AI regulation, modelled in broad architecture on the European Union AI Act then advancing through the EU legislative process, but with a more explicit fundamental-rights and human-dignity framing and a stronger integration with Brazil's General Data Protection Law (LGPD) than the EU text. The Senate's Temporary Internal Commission on Artificial Intelligence (CTIA) — Comissão Temporária Interna sobre Inteligência Artificial — was constituted to handle the bill, with its first meeting held on 16 August 2023; Senator Carlos Viana was elected CTIA president and Senator Astronauta Marcos Pontes vice-president.

The Coalizão Direitos na Rede — whose roughly 50 member organisations include the full field of Brazilian digital-rights civil society active on AI and platform-regulation files — moved quickly: a letter of support for PL 2338/2023 was issued on 14 June 2023, endorsing the framework while flagging the need for strengthened high-risk prohibitions and governance mechanisms. The coalition's position was that the Pacheco bill represented an important legislative opening — the first federal AI framework proposal advanced by a Senate president — but that without civil society pressure it risked being weakened through industry lobbying before reaching the Senate floor.

Civil society positions and the CDR technical note

On 23 August 2023, the Coalizão Direitos na Rede published its substantive technical note (nota técnica) on PL 2338/2023, the most comprehensive civil-society analytical document produced during the Senate hearing phase. The note's core demands were:

• Categorical prohibitions on high-risk AI uses: facial recognition in public-security contexts; predictive-policing systems; social-scoring of citizens; emotion-recognition in employment and educational settings; autonomous weapons systems. The CDR position was that these uses should be categorically prohibited by statute rather than classified as high-risk and left to case-by-case governance mechanisms — the same demand Intervozes and Coding Rights would carry into the 2024 open letter on FRT in the federal bill.
• Risk-based classification with a strong high-risk tier: AI systems in health, education, employment, housing, social services, and justice should be classified as high-risk and subject to mandatory fundamental-rights impact assessments, human-oversight requirements, and independent auditing before deployment.
• An independent national AI authority: the CDR argued against assigning AI oversight to existing sectoral regulators or to a body housed in the executive branch with weak institutional independence, and for a new authority with genuine sanctioning power and civil-society representation in governance.
• Worker protections: mandatory disclosure to workers of AI-system use in algorithmic management, hiring, dismissal, and performance monitoring; rights to contest automated decisions; collective bargaining on AI deployment conditions. This was a pressure point where the coalition would later record ground lost to industry lobbying in the rapporteur phase.
• Transparency and explainability obligations: public registries of high-risk AI deployments; mandatory explainability to affected individuals of consequential automated decisions; researcher and regulator access to AI systems under audit.

Two days later, on 25 August 2023, LAPIN — the Laboratório de Políticas Públicas e Internet — published its own independent technical note on PL 2338/2023, analysing the bill's risk-classification framework, governance architecture, and liability regime from a policy-research perspective that overlapped with but was distinct from the CDR coalition position. The pair of technical notes in the final week of August 2023 established the substantive civil-society record that the CTIA would subsequently draw on in its hearing programme.

The October 2023 hearing series

The CTIA concentrated its major public-hearing programme into a series of five sessions between 24 and 31 October 2023. The hearing schedule covered:

• 24 October (9th CTIA meeting): National and international regulatory approaches — including participation by ANPD (Brazil's data-protection authority) Director Miriam Wimmer and Anatel Superintendent Abraão Balbino e Silva, establishing the regulatory-architecture and inspection-regime questions that the civil-society position on an independent AI authority was pitched against.
• 25 October (10th meeting): Risks, development, innovation, and labour-market effects — including the presence of a federal judge (Caroline Somesom Tauk) and the internet governance council (NIC.br Director Demi Getschko) alongside labour-market researchers.
• 26 October (11th meeting): Central aspects of AI regulation — including Carlos Affonso de Souza, Marcela Mattiuzzo, and others from the Brazilian tech-law and civil-society academic field, covering the principiological approach, liability regime, multi-sectoral governance, automated-decisions rights, and human oversight.
• 31 October AM (12th meeting): AI in elections and information dissemination — including Renata Mielli, coordinator of CGI.br (the Brazilian internet steering committee), whose presence brought the internet-governance community's perspective to bear on the AI-in-democracy questions the CDR's broader member field (including Intervozes on platform regulation and Coding Rights on feminist-tech methodology) had been working since PL 2630/2020.
• 31 October PM (13th meeting): AI in cybercrime and the criminal-justice system.

Before the hearings opened, the CDR published a pre-hearing open letter on 20 October 2023 — "Senadores, precisamos falar sobre Inteligência Artificial" ("Senators, we need to talk about Artificial Intelligence") — pressing the CTIA to ensure that the hearing witness lists included representatives of the communities most affected by AI systems: Black Brazilians, platform and gig workers, women, Indigenous peoples, and LGBTQIA+ people. The coalition documented that the initial hearing programme over-indexed on industry, regulatory-body, and technical-expert witnesses relative to grassroots and affected-community civil society, and identified Paula Guedes — CDR AI Working Group researcher and member of the Legalite Nucleus at PUC-RJ — as its principal hearing representative for the sessions where civil society did participate.

Significance and later arc

The 2023 hearing phase of PL 2338/2023 is the corpus's first recorded instance of Brazil's digital-rights civil-society field engaging a federal AI-governance legislative process in a structured, coalition-coordinated form. The Coalizão Direitos na Rede's threefold engagement — a support letter establishing early endorsement of the framework, a substantive technical note anchoring the coalition's demands, and a pre-hearing advocacy push for representative witness lists — maps the strategic posture Brazilian civil society would maintain through the bill's multi-year legislative arc: broadly supportive of the risk-based framework while consistently pressing for stronger high-risk prohibitions, independent governance, and worker protections than the Senate text as filed carried.

In the period between the 2023 hearing phase and the December 2024 Senate floor vote, civil society recorded substantial gains and some losses. The July 2024 CDR open letter calling for facial-recognition prohibitions in PL 2338/2023 — signed by over 50 civil society and academic organisations including Coding Rights and Intervozes — was a direct escalation of the 2023 hearing-phase demand, with the CDR documenting that worker-protection provisions had been weakened through industry lobbying in the rapporteur phase. The substitute report by rapporteur Senator Eduardo Gomes (PL-TO) was approved by the CTIA on 5 December 2024; the Senate floor approved PL 2338/2023 on 10 December 2024 and transmitted the bill to the Chamber of Deputies on 17 March 2025, where the civil-society engagement arc continues.

For the corpus's wider movement-graph frame, the Brazil PL 2338/2023 hearing process belongs to the same register as the EU AI Act advocacy by Access Now and allied European civil-society organisations (which culminated in the March 2024 European Parliament vote elsewhere in the corpus): domestic civil-society coalitions intervening in national AI-governance legislation, using the public-hearing and technical-note mechanisms of representative democracy to route grassroots and affected-community demands into legal text, at a moment when most advanced economies were drafting their first AI frameworks. Brazil's version of that engagement is distinctive in two respects: it is rooted in an older direito à comunicação / media-democracy tradition that gives organisations like Intervozes a prior-decade vocabulary of democratic platform regulation to extend into the AI file; and it is animated by a specifically Brazilian analysis of risk — the country's racialized criminal-justice system, the scale of gig-economy platform deployment, and the structural conditions of Brazil's three-hundred-million-person digital public — that the CDR member field (spanning Coding Rights's transfeminist methodology, Intervozes's collective-media-democracy frame, ARTICLE 19 Brasil's free-expression tradition, and others) brings to the federal legislative table.