Technical Note 12/2025 of Brazil's data protection authority, the ANPD, dated 3 April 2025, consolidates the 124 contributions to its consultation on AI and the review of automated decisions, held from 6 November 2024 to 24 January 2025. The consultation serves item 7 of the ANPD's 2025-2026 regulatory agenda, which includes regulating the data subject's right under Article 20 of the LGPD to request review of automated decisions. Credit decisions recur in the contributions as the standard example of a high-impact automated decision. The note is in Portuguese and is not binding.
OFFICIAL TEXT: gov.br ↗ · FINAL · BRAZIL (BCB / CMN / ANPD)
| Document | ANPD Technical Note 12/2025 — Nota Técnica nº 12/2025/CON1/CGN/ANPD: consolidation of contributions to the consultation on artificial intelligence and review of automated decisions |
| Issued by | Banco Central do Brasil (BCB) and the National Monetary Council (CMN), with the National Data Protection Authority (ANPD) and the National Congress (AI bill PL 2338/2023) |
| Type | Consultation |
| Status | Final |
| Published | Apr 3, 2025 |
| Applies to | Controllers processing personal data under the LGPD, including banks that make automated decisions such as credit decisions. A consultation summary: the ANPD's own rules state that a consultation does not represent its final position |
| Official source | gov.br ↗ |
| Use cases | Data & privacy · Credit scoring & underwriting · Fair lending & discrimination · Generative & agentic AI |
What are the key points of ANPD Technical Note 12/2025?
- The consultation ran from 6 November 2024 to 5 December 2024 and was extended to 24 January 2025; it had 15 questions in five blocks, starting with LGPD principles, legal bases and data subject rights.
- 124 participants: 99 through the Participa Mais Brasil platform and 25 by email; about 56% answered on behalf of a data processing agent.
- The project regulates item 7 (artificial intelligence) of the ANPD's 2025-2026 regulatory agenda, approved by Resolution 23 of 9 December 2024, in particular Article 20 of the LGPD on review of automated decisions.
- The ANPD says it may also address the topic through guidance documents such as guides and technical studies.
- Views on human review diverged: some contributions called for escalation to human review in complex cases or when a decision is contested; one said any automated decision affecting an individual should be reviewed by a human, while others said only high-impact decisions such as granting credit, approving a loan or setting a risk profile should be.
- There was no consensus on whether generative AI models fall within the review criteria of Article 20.
- The note concludes that it should serve as input to the regulatory project and be sent to the ANPD's Board of Directors.
What did ANPD Technical Note 12/2025 change for banks?
The note records the evidence base for the ANPD's first rules on AI and automated decisions under the LGPD. No regulation had been issued under item 7 by the time of this review.
What is ANPD Technical Note 12/2025?
It is the ANPD's consolidation of the 124 contributions to its consultation on artificial intelligence and the review of automated decisions, held from November 2024 to January 2025. Dated 3 April 2025, it feeds the regulation of Article 20 of the LGPD under item 7 of the ANPD's 2025-2026 regulatory agenda.
Does the ANPD require human review of automated credit decisions?
Not yet by specific rule. Technical Note 12/2025 summarises views, some of which say high-impact decisions such as granting credit or approving a loan should be reviewed by a human, but the ANPD's own procedure states that a consultation does not represent its final position.
| Date | Document | Status |
|---|---|---|
| Nov 11, 2025 | BCB survey on AI in the National Financial System (FSR, Nov 2025) — Survey on the Use of Artificial Intelligence in the National Financial System (Financial Stability Report, Volume 24, No. 2, November 2025, selected topic 2.1) | Final |
| Mar 17, 2025 | PL 2338/2023 (Brazil AI bill) — Projeto de Lei nº 2.338, de 2023: development, promotion and ethical and responsible use of artificial intelligence based on the centrality of the human person | Proposed |
| Feb 26, 2021 | CMN Resolution 4,893 — Resolution CMN 4,893 of February 26, 2021: cyber security policy and requirements for contracting data processing, data storage and cloud computing services | In force |
| Nov 19, 2026 | BNM AI Feedback Statement — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial Sector | Final |
| Nov 4, 2026 | OCC Bulletin 2026-54 — Super Intelligence: Supervisory Framework | Final |
| Oct 15, 2026 | FSB AI sound practices (final report, Oct 2026) — Sound Practices for Responsible Adoption of Artificial Intelligence (AI): Final report | Final |
Follow every document these regulators publish
when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning