Record STMT-W62D261Voluntary

Policy on the Use of Artificial Intelligence in the Administration of Justice

POLICY ON THE USE OF ARTIFICIAL INTELLIGENCE IN THE ADMINISTRATION OF JUSTICE

Institutional policy Institutional policySouthern Europe · ESdraft

This document sets out Spain's preliminary policy on the use of artificial intelligence within the Administration of Justice. It establishes minimum criteria to ensure responsible, legal, and ethical use of AI tools, differentiating between tools that affect jurisdictional activities and those that do not. The policy covers guiding principles, responsibilities, definitions, and specific tables of permitted, restricted, and prohibited use cases.

Summary and content fields are drawn from the document text and are pending editorial review.

Who issued it

Gobierno de EspañaGovernment (executive) · ES

Lead issuer

Legal forceVoluntary

Under whose authorityInstitutional policy

Judged fromfull text

Who it governs

Public sector bodiesCompaniesAI deployers and usersProfessional group

This policy is for all employees of the Administration of Justice, legal professionals, AI tool providers, and institutional actors who access or manage judicial data in Spain.

Sectors

Justice · Administration of JusticePublic administration

What the document contains

What's covered

The policy addresses the scope, guiding principles, operational rules, responsibilities, definitions, and specific use-case classifications for artificial intelligence in judicial administration. It details what types of AI uses are permitted internally, which require formal management or IT approval, and which are strictly prohibited.

What's missing

As a preliminary version, it notes that it may contain translation errors and does not provide final legislative enforcement mechanisms or fixed statutory penalties for violations outside of administrative and judicial oversight frameworks.

Key takeaways

AI systems in the Administration of Justice must fully respect fundamental rights, data protection regulations, and judicial independence.
Human review is mandatory for any AI-generated output that directly or indirectly affects user rights or jurisdictional activities.
The General Council of the Judiciary (CGPJ) oversees algorithmic surveillance for systems affecting judicial independence.
Personal or confidential judicial data cannot be used in public or unapproved AI tools.
Use cases are strictly categorized into permitted, requiring management approval, requiring IT approval, and prohibited.

Principles it states

Principle of respect for fundamental rightsPrinciple of non-discriminationPrinciple of quality and securityPrinciple of transparency, impartiality and fairnessPrinciples under user controlPrinciple of Equity and Universal AccessPrinciple of Prevention of Bias and DiscriminationPrivacy and Personal Data Protection PrinciplePrinciple of Responsible Innovation and Continuous EvaluationTraining PrincipleCo-governance Principle

Outline

· 1.1 AIM
· 1.2 AREA OF APPLICATION
· 1.3 RECIPIENTS
· 1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY
· 1.5 RULES FOR THE USE OF ARTIFICIAL INTELLIGENCE SYSTEMS
· 1.6 AI RESPONSIBILITIES AND CONTROL
· ANNEX I: DEFINITIONS
· ANNEX II: PERMITTED AND PROHIBITED USE CASES

Issued2024

Length3,678 words · 12 pages

FormatNot recorded

Scopenational

How it scores

Pending review

Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.

Accountability

4/5

Clear responsibility for AI outcomes

Regarding the use of AI : the person responsible for its use will be the court, tribunal, judicial office or prosecutor’s office, and within each of them, the person who makes effective use of it.

1.6 AI RESPONSIBILITIES AND CONTROL

screening score · 2026-10-23

Fairness

4/5

Non-discrimination and equitable treatment

It is essential to carry out periodic evaluations of the algorithms to identify and correct possible biases inherent in the training data or algorithms used.

1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY

screening score · 2026-10-23

Human Oversight

5/5

Humans remain in control of consequential decisions

Furthermore, AI should never replace human decision-making in crucial matters of the Administration of Justice. It is essential that AI systems are used as tools to assist legal professionals in their duties, but the ultimate responsibility for making legal decisions must rest on Judges and Magistrates.

1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY

screening score · 2026-10-23

Privacy

4/5

Protection of personal data and private life

Strict privacy and personal data protection measures must be established in the use of AI in the Administration of Justice, including the need to previously carry out data protection impact assessments.

1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY

screening score · 2026-10-23

Transparency

4/5

Openness about how AI systems work and are used

Specifically, it is desirable to publish FAT (Fairness, Accuracy and Transparency) records or similar instruments, about the data used, members of the AI teams, services, algorithms, possible biases and applications

1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY

screening score · 2026-10-23

Enforcement

4/5

Consequences and compliance mechanisms with real force

AI applications that have not been authorised should not be used. 6. Unapproved application programming interfaces (APIs), add-ons, connectors, or software related to AI systems, should not be installed.

1.5 RULES FOR THE USE OF ARTIFICIAL INTELLIGENCE SYSTEMS

screening score · 2026-10-23

Impact Assessment

4/5

Required assessment of an AI system's risks/impacts before or during use

Strict privacy and personal data protection measures must be established in the use of AI in the Administration of Justice, including the need to previously carry out data protection impact assessments.

1.4 BASIC PRINCIPLES THAT GUIDE THIS USE POLICY

screening score · 2026-10-23

Topics

artificial_intelligence justice_administration judicial_independence ethics data_protection algorithm_transparency