The Bottom Line
- Human Oversight is Non-Negotiable: For businesses in litigation, this means legal arguments must continue to persuade a human judge. While Spanish courts may use AI for back-office efficiency, the final decision-making, weighing of evidence, and legal interpretation remain exclusively human tasks.
- A “Walled Garden” for Legal Tech: AI vendors targeting the Spanish justice system face a clear approval process. Only tools vetted and supplied by official judicial bodies will be permitted, creating a high bar for entry focused on security, transparency, and bias prevention.
- Strict Prohibitions on High-Risk AI: The guidelines explicitly forbid using AI for profiling individuals, predicting behavior, or risk assessments. This provides assurance that judicial AI will not be used for “predictive justice,” maintaining a level playing field for corporate litigants.
The Details
In a significant move to regulate the use of emerging technology within its justice system, Spain’s General Council of the Judiciary (CGPJ) has issued formal instructions for judges and magistrates on the use of artificial intelligence. The guidance aims to create a clear and consistent framework, aligning with both Spanish and EU regulations like the recent AI Act. The core objective is to harness the efficiency of AI for support tasks without compromising the foundational principles of judicial independence, responsibility, and the fundamental rights of individuals.
The cornerstone of the new rules is the principle of “effective human control.” The instructions firmly position AI as an assistant, never a substitute for a judge. AI tools cannot operate autonomously or be used to delegate core judicial functions such as assessing facts, evaluating evidence, or applying the law. While judges can use approved AI for tasks like legal research, summarizing documents, or structuring information, the ultimate intellectual work and the legal responsibility for any judgment rest solely with the human judge. Even AI-generated drafts of rulings require a “critical, complete, and personal” review and validation by the magistrate before any part can be adopted.
The guidelines also establish a controlled environment for AI adoption. Judges may only use applications provided and vetted by competent justice administrations or the CGPJ itself, effectively creating a “walled garden” to ensure security and reliability. The instructions draw explicit red lines, prohibiting the use of AI for any form of human profiling, behavioral prediction, or risk classification. Furthermore, AI systems cannot be used to process specially protected personal data, reinforcing strict data confidentiality and preventing the technology from making judgments based on sensitive information.
Source
Consejo General del Poder Judicial (CGPJ)
