safelegalaidata/justice-ai-deployments

Justice-sector AI deployments and procurement

595 official records of AI used by courts, judiciaries, prosecutors, public defenders, legal-aid bodies, tribunals and justice ministries worldwide — use-case inventory entries, procurement notices and awards, court announcements, judgments in which a court discloses its own AI use, and judicial-conduct opinions on judges' AI use.

Built 2026-09-08 by SafeLegalAI (Cognesio LLP). Canonical pages: safelegalai.com/public-sector · repository, pipeline and issues: https://github.com/SafeLegalAI/justice-ai-deployments · this mirror: https://huggingface.co/datasets/safelegalaidata/justice-ai-deployments.

table	rows	one row is
deployments	595	one official record of AI used by a justice body

Every row carries source_url, fetched_at and, where the Wayback Machine accepted the page, archive_url; url links the canonical page on safelegalai.com; notice carries the terms below. Full schemas: schema/.

One record, one row: body, body_type, country, region, record_type (inventory-entry · procurement-notice · contract-award · announcement · pilot-report · judgment-disclosure · conduct-opinion · policy · survey-entry), tool_or_use_case, vendor, ai_kind, status_as_stated (the source's own word), status_date, value/currency as published, a ≤50-word summary, a ≤25-word quote, licence, and provenance.

deployments by country

value	rows
US	412
BR	98
NL	31
GB	21
EE	5
IN	5
ES	5
AE	4
CN	3
SG	2
NO	2
CO	2
DE	1
HK	1
EU	1

deployments by body_type

value	rows
police-forensics	177
tribunal	79
court	69
prosecutor	69
other-agency	58
corrections-probation	56
justice-ministry	41
judiciary	22
court-administration	21
legal-aid	3

deployments by record_type

value	rows
inventory-entry	412
survey-entry	92
announcement	44
policy	25
transparency-record	9
judgment-disclosure	4
contract-award	3
pilot-report	3
conduct-opinion	2
procurement-notice	1

Method

Machine-readable official inventories first (US DOJ AI Use Case Inventory, UK Algorithmic Transparency Records via the gov.uk Search API, Brazil CNJ), then procurement portals (SAM.gov, USAspending, Contracts Finder OCDS, TED) filtered to justice bodies, then court and judiciary announcements, judgments disclosing judicial AI use, and judicial-conduct advisory opinions. Status words are copied from the source, never upgraded from pilot to deployed. CEPEJ, NCSC and the AI Incident Database were used as leads only and every row was re-sourced to an official page.

SafeLegalAI records what courts, regulators, legislatures and vendors' own public pages state; it does not infer, rank or advise. Coding columns are SafeLegalAI's good-faith reading for comparison, not findings about any person or body. Corrections and right of reply: safelegalai.com/report.

Licence and notices

US federal records are public domain (17 U.S.C. § 105); UK records are OGL v3; EU/TED under Decision 2011/833/EU; Brazilian official acts under Lei 9.610 art. 8; court opinions public domain or quoted at ≤25 words. The compilation and coding are CC BY 4.0 — attribute SafeLegalAI (safelegalai.com), published by Cognesio LLP. Code is Apache-2.0.

Provided as is, without warranty. Not legal advice. SafeLegalAI (Cognesio LLP) records what courts, regulators, legislatures and vendors' own public pages state; the linked official documents are the record. Names and marks belong to their owners. Anyone named may reply: https://safelegalai.com/report. Full terms: https://safelegalai.com/disclaimer See DISCLAIMER.md and NOTICE in this repository.

Uses

Suited to: counting and comparing what the record shows (by court, jurisdiction, date, actor, outcome, status); building watch-lists and alerts from source_url/fetched_at; grounding retrieval or summarisation on cited primary documents; teaching and library guides that need a dated, sourced list.

Not suited to: ranking products, people or courts; inferring prevalence beyond what a court or regulator has itself stated; any use that treats a coding column as a finding of fact or law. Where a row names a person or organisation it does so as they appear in a public document; anyone named may request a correction or right of reply at https://safelegalai.com/report.

Cite

SafeLegalAI (Cognesio LLP), "Justice-sector AI deployments and procurement", v0.1.1, 2026-09-08. https://huggingface.co/datasets/safelegalaidata/justice-ai-deployments — CC BY 4.0. Canonical: https://safelegalai.com/public-sector

@dataset{safelegalai_justice_ai_deployments_0_1_1,
  title        = {Justice-sector AI deployments and procurement},
  author       = {{SafeLegalAI (Cognesio LLP)}},
  year         = {2026},
  version      = {0.1.1},
  url          = {https://safelegalai.com/public-sector},
  note         = {Mirror: https://huggingface.co/datasets/safelegalaidata/justice-ai-deployments. Data CC BY 4.0. Built 2026-09-08.}
}