Argentina - Río Negro - AI Good Practices Protocol (RAI-AR-RN-PDBPPXX-2024)

Overview

The 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' (Protocol of Good Practices for the Use of Generative Artificial Intelligence), identified as argentina-2024-rio-negro-ai-judicial-protocol, represents a landmark regulatory initiative from the Superior Tribunal de Justicia (STJ) of the Province of Río Negro, Argentina. Approved initially through Acordada 15/2024, this document sets forth a comprehensive framework of guidelines and recommendations for the ethical, responsible, and diligent integration of generative AI (IAGen) tools within the judicial system. Its primary aim is to leverage the efficiency-enhancing capabilities of AI while rigorously upholding the foundational principles of justice, human oversight, and data protection. The protocol explicitly addresses the use of large language models such as ChatGPT, Gemini, Copilot, and Claude, recognizing their potential to support judicial functions but also acknowledging the critical need for robust safeguards.

This pioneering guideline underscores the commitment of the Río Negro judiciary to navigating the transformative landscape of artificial intelligence with prudence and foresight. It emphasizes that AI tools are to serve as complements to human judgment, not substitutes, ensuring that the ultimate responsibility for judicial decisions remains firmly with magistrates and officials. The protocol's scope was further broadened by Acordada 22/2025, extending its application to external auxiliaries, including lawyers and experts, thereby encompassing a wider array of actors within the judicial ecosystem. This expansion reflects an adaptive regulatory approach, acknowledging the pervasive nature of AI technologies and the necessity for consistent ethical standards across all facets of legal practice. The introduction of potential sanctions for misuse further solidifies the protocol's intent to enforce responsible AI adoption and maintain public trust in the administration of justice.

Definitions

The Protocol of Good Practices for Generative AI in Río Negro's judiciary implicitly and explicitly defines several key terms crucial for its understanding and application. 'Inteligencia Artificial Generativa (IAGen)' refers to advanced AI systems, such as large language models like ChatGPT, Gemini, Copilot, and Claude, that are capable of producing novel content, including text, code, or other data, based on patterns and structures learned from vast datasets. These tools are recognized for their potential to assist in tasks like ideation, analysis, drafting, and content correction within the judicial context.

Critical concepts addressed within the protocol include 'Alucinaciones' (Hallucinations) and 'Sesgos' (Biases). Hallucinations denote instances where IAGen systems generate information that is factually incorrect, fabricated, or lacks real-world corroboration, despite being presented as authoritative. Biases, on the other hand, refer to systemic prejudices or unfair inclinations embedded within AI outputs, often stemming from skewed or incomplete training data, which can lead to discriminatory or inequitable outcomes. The protocol also highlights 'Control Humano Suficiente' (Sufficient Human Control), emphasizing the indispensable role of human critical judgment, verification, and ultimate decision-making, ensuring that AI remains a tool under human direction. 'Responsabilidad Funcional y Ética' (Functional and Ethical Responsibility) reinforces that legal and ethical accountability for judicial acts cannot be delegated to AI systems, resting solely with the human operators.

Governance and Institutional Framework

The governance and institutional framework for the 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' is firmly anchored within the Superior Tribunal de Justicia (STJ) of the Province of Río Negro. The STJ, as the highest judicial body in the province, holds the ultimate authority for the approval, oversight, and continuous adaptation of this protocol. Its initial enactment through Acordada 15/2024 demonstrates the STJ's proactive stance in addressing the implications of AI within the judiciary. The protocol mandates that any projects involving the use of IAGen in judicial workflows must receive prior authorization from the STJ, ensuring centralized control and alignment with institutional objectives. Furthermore, the Presidency of the Committee for the Computerization of Judicial Management, following a review by the General Directorate of Systems, is responsible for authorizing the personal use of generative AI by magistrates, officials, and employees for their assigned functions.

Beyond initial approval and project authorization, the institutional framework extends to ongoing support and enforcement mechanisms. The protocol entrusts the Judicial Training School, in collaboration with the General Directorate of Systems, with the crucial task of developing and implementing a mandatory training plan for all judicial personnel regarding the use of IAGen. This commitment to continuous education is vital for fostering an informed and responsible user base. Moreover, the subsequent Acordada 22/2025, which extended the protocol's reach to external auxiliaries, also introduced provisions for sanctions in cases of misuse. This highlights the STJ's role not only in setting guidelines but also in establishing accountability and corrective measures, reinforcing the importance of compliance across the entire judicial ecosystem of Río Negro. The STJ's oversight ensures that the integration of AI remains consistent with legal principles, ethical standards, and the imperative of maintaining public trust in the judicial process.

Key Focus Areas

The 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' in Río Negro is meticulously structured around several key focus areas designed to ensure the responsible and ethical integration of AI within the judicial system. A paramount focus is on governance_oversight, establishing clear institutional responsibilities and authorization processes. This includes the requirement for all AI-driven projects to be pre-approved by the Superior Tribunal de Justicia (STJ) and for individual use of IAGen to be authorized by the Committee for the Computerization of Judicial Management. This centralized oversight aims to prevent uncoordinated or potentially risky deployments of AI, ensuring that all applications align with the judiciary's strategic goals and ethical standards. The protocol also mandates continuous training programs, underscoring a commitment to equipping personnel with the necessary knowledge to use AI tools effectively and responsibly.

Another critical area of emphasis is transparency_disclosure, particularly concerning the explainability and verification of AI-generated content. The protocol explicitly states that the use of IAGen tools does not diminish the functional and ethical responsibilities of judicial operators. It stresses the imperative of 'sufficient human control,' meaning that all outputs from AI systems must be rigorously validated, critically analyzed, and cross-referenced with reliable sources and professional legal knowledge before being incorporated into judicial acts. This directive directly addresses the risk of AI 'hallucinations' or the generation of misleading information, ensuring that human judgment remains the ultimate arbiter of truth and accuracy. The protocol also implicitly supports fundamental_rights by aiming to mitigate risk_management associated with biases and discrimination, particularly in the context of personal data protection, which is a core concern when AI processes information. The extension of the protocol to external auxiliaries further broadens the scope of these transparency and verification requirements, promoting a culture of accountability across all judicial stakeholders.

Implementation Framework

The implementation framework for the 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' in Río Negro is designed to ensure a structured and controlled integration of AI technologies across the judicial branch. The foundational element is the requirement for prior authorization for any project involving IAGen, which must be submitted to and approved by the Superior Tribunal de Justicia (STJ). This centralized approval process ensures that all significant AI deployments are strategically aligned with the judiciary's objectives and undergo a thorough ethical and technical review. For individual use of IAGen tools by magistrates, officials, and employees in their daily tasks, authorization is required from the Presidency of the Committee for the Computerization of Judicial Management, following a technical assessment by the General Directorate of Systems. This tiered authorization system balances the need for innovation with stringent oversight.

A critical component of the implementation framework is the mandatory training and capacity-building initiative. The protocol explicitly tasks the Judicial Training School, in collaboration with the General Directorate of Systems, with developing and delivering an obligatory training plan for all personnel on the responsible use of IAGen. This continuous education is essential for fostering a deep understanding of AI functionalities, limitations, potential impacts, and associated risks, thereby promoting informed and ethical usage. Furthermore, the protocol outlines specific directives for data handling, emphasizing the protection of personal data and, where necessary, its anonymization before processing with AI tools. The subsequent expansion of the protocol's scope to external auxiliaries through Acordada 22/2025 signifies an evolving framework that adapts to the broader ecosystem of justice administration, ensuring that consistent standards of AI use and accountability apply to all parties interacting with the judicial system. This adaptability is crucial for maintaining the protocol's relevance and effectiveness in a rapidly advancing technological landscape.

Monitoring and Evaluation

The 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' in Río Negro implicitly establishes mechanisms for monitoring and evaluation, primarily through its emphasis on human oversight, continuous review, and the potential for sanctions. While the document itself, as a protocol of good practices, may not detail a formal, standalone monitoring body, the requirement for human validation of all AI-generated content serves as an ongoing, decentralized monitoring process. Magistrates, officials, and now external auxiliaries are individually responsible for critically reviewing, verifying, and confirming the accuracy and relevance of any information or drafts produced by IAGen tools. This continuous human intervention acts as a primary safeguard against errors, biases, or 'hallucinations' from AI systems, effectively performing real-time quality control.

Furthermore, the directive for the protocol to be "revisado y actualizado constantemente" (constantly reviewed and updated) by the Superior Tribunal de Justicia (STJ) indicates an institutional commitment to ongoing evaluation. This implies that the STJ, likely through its Committee for the Computerization of Judicial Management and the General Directorate of Systems, will regularly assess the protocol's effectiveness, identify emerging challenges, and incorporate necessary adjustments. The detection of instances where AI generated false citations or jurisprudence, as highlighted in the context of Acordada 22/2025, serves as a clear example of how misuse or shortcomings are identified and lead to amendments and stricter enforcement. This reactive and adaptive approach, coupled with mandatory training and the potential for corrective actions and sanctions, forms a comprehensive, albeit evolving, system for monitoring and evaluating the protocol's implementation and impact within the Río Negro judiciary.

Penalties, Liability, and Appeals

While the initial Acordada 15/2024 establishing the 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' primarily focused on guidelines and recommendations, the subsequent Acordada 22/2025 introduced explicit provisions for penalties and liability, particularly extending to external auxiliaries. This amendment was prompted by concrete instances of misuse, such as lawyers presenting legal documents containing fabricated citations or non-existent jurisprudence generated by AI. The Superior Tribunal de Justicia (STJ) recognized the severe implications of such practices for the integrity of judicial processes and the right to defense. Consequently, the updated framework enables tribunals to apply sanctions within their corrective powers, as defined by the Organic Law and existing procedural codes.

Crucially, Acordada 22/2025 also mandates that cases of improper use or non-compliance with the protocol by external auxiliaries, such as lawyers, must be referred to their respective professional colleges for consideration by Ethics or Disciplinary Tribunals. This establishes a dual layer of accountability, encompassing both judicial administrative sanctions and professional ethical oversight. The protocol unequivocally states that the use of IAGen does not diminish the functional and ethical responsibilities of any judicial operator or auxiliary, meaning that ultimate liability for the content and accuracy of submissions rests with the human professional. While specific appeal mechanisms for these new sanctions are not detailed within the public summaries of the acordadas, it can be inferred that appeals would follow established judicial and professional disciplinary procedures, ensuring due process for those subject to corrective actions.

Relationship to Other Instruments

The 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' in Río Negro operates within and complements a broader legal and ethical landscape. It explicitly references and aligns with the existing constitutional and legal responsibilities of magistrates, officials, and employees, underscoring that the introduction of AI tools does not diminish these fundamental duties. The protocol is also framed within the context of the informatization and digitalization of judicial management systems, which are viewed as strategic allies for enhancing efficiency and agility. This suggests an integration with ongoing digital transformation initiatives within the Poder Judicial de Río Negro, aiming for a cohesive technological advancement.

Furthermore, the protocol draws inspiration and guidance from international and national best practices in AI regulation. It is notably based on the 'Guía de inteligencia artificial para los funcionarios judiciales' (AI Guide for Judicial Officials) from the United Kingdom's judiciary, approved in December 2023, and the 'Guía de directrices para el uso de ChatGPT e IA generativa de texto en la Justicia' (Guidelines for the Use of ChatGPT and Generative Text AI in Justice) from the UBA IALAB research team (University of Buenos Aires, Argentina, 2023). This demonstrates a commitment to incorporating global and local expertise in shaping its directives. The subsequent Acordada 22/2025, which extended the protocol's scope and introduced sanctions, also references the Organic Law 5731 (Ley Orgánica del Poder Judicial) and the "Normas de Ética Profesional" (Professional Ethics Standards) by the Federación Argentina de Colegios de Abogados, ensuring consistency with established legal and ethical frameworks governing the conduct of external auxiliaries like lawyers and experts.

International Alignment

The 'Protocolo de Buenas Prácticas para el uso de Inteligencia Artificial Generativa' from Río Negro, Argentina, demonstrates a clear commitment to international alignment by drawing upon established global and national best practices in AI governance. The protocol explicitly states that its development was informed by the 'Guía de inteligencia artificial para los funcionarios judiciales' (AI Guide for Judicial Officials) from the United Kingdom's judiciary, which was approved in December 2023. This direct reference signifies an effort to incorporate insights and frameworks from other advanced jurisdictions that have already begun to grapple with the ethical and practical challenges of AI in legal contexts. Such cross-jurisdictional referencing helps ensure that the Río Negro protocol is built upon a foundation of shared principles and lessons learned from international experiences.

In addition to international guidance, the protocol also aligns with national academic and research initiatives, specifically citing the 'Guía de directrices para el uso de ChatGPT e IA generativa de texto en la Justicia' (Guidelines for the Use of ChatGPT and Generative Text AI in Justice) developed by the UBA IALAB research team from the University of Buenos Aires, Argentina, in 2023. This dual approach of referencing both international judicial guidelines and prominent national academic research underscores a comprehensive strategy for developing a robust and informed regulatory instrument. By integrating these diverse sources, the Río Negro protocol aims to ensure that its good practices are not only locally relevant but also globally informed, contributing to a broader international discourse on responsible AI in the justice sector and potentially facilitating future cross-border cooperation or harmonization efforts in AI regulation.

Implementation Timeline

Sources and References