evidoria

← Back to browse

Good practice Imported

Sinapses — Brazil's Mandatory National Registry and Risk-Audit Platform for Judicial AI

Brazil · Brasília · See the Brazil profile · See the Brasília profile

Evidence: Descriptive / self-reported Top 6% 87/100 · Ask Evidence Copilot about this practice

Brazil's National Council of Justice runs Sinapses, a national catalogue where all AI models used or developed by the Brazilian judiciary must be registered, risk-classified and — for high-risk systems — subjected to periodic audits, with about 150 active models from 29 courts an

~150
Active AI models registered (2025)
29
Courts and councils contributing (2025)
Sinapses — Brazil's Mandatory National Registry and Risk-Audit Platform for Judicial AI

Details

Maturity
Established
Promoter
Conselho Nacional de Justiça (CNJ), built with Tribunal de Justiça de Rondônia (TJRO) and UNDP Brazil (Justiça 4.0 programme)
Period
2020–present (mandatory since 2025)
Keywords
judicial AI, algorithmic transparency, AI governance, public administration

Context

Sinapses is a national platform run by Brazil's Conselho Nacional de Justica (CNJ), originally built by the Tribunal de Justica de Rondonia and expanded nationwide with the UNDP-backed Justica 4.0 programme. First established by CNJ Resolucao n. 332/2020, it became a binding governance requirement under Resolucao CNJ n. 615/2025 (in force since 14 July 2025).

Activities

Every AI solution developed or used by a Brazilian court must be registered in Sinapses and risk-classified; high-risk systems must additionally publish a public summary of their algorithmic impact assessment and undergo periodic audits.

Results

As of 2025, Sinapses lists roughly 150 active AI models contributed by 29 courts and councils across the federal and state judiciary -- the first system-level, judiciary-wide inventory of its kind in Brazil.

Conclusions

The mandate is judiciary-specific and does not cover Brazil's executive-branch agencies, which use separate, less standardised AI oversight. The depth and rigour of individual courts' risk classifications has not yet been independently audited by a third party, so registration completeness and quality may vary across the 29 participating bodies.

Implementation

Indicative cost
Medium (€50k–€500k) — Platform built with UNDP technical support under the Justica 4.0 programme; no separate budget figure disclosed.
Time to results
Long (> 3 years) — Established 2020; mandatory since 14 July 2025.
Staffing & skills
Conselho Nacional de Justica (CNJ), Tribunal de Justica de Rondonia (TJRO), UNDP Brazil (Justica 4.0 programme), participating courts' AI governance units

Conditions for success

  • binding resolution (CNJ 615/2025) making registration mandatory rather than voluntary
  • public disclosure requirement for high-risk systems' impact assessments
  • periodic audit obligation for high-risk systems

Common failure modes

  • quality and depth of individual courts' self-reported risk classifications not yet independently audited

Where it fits

Governance type
national judicial council, binding mandate
Scale
national judiciary (Brazil)
Income level
upper-middle income (Brazil)

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful