Buenos Aires AI Guidelines for the Judiciary

Overview

The City of Buenos Aires Presidency Resolution No. 206/2025, specifically its Annex I, establishes a comprehensive set of recommendations and guidelines for the ethical, responsible, and sustainable use of Artificial Intelligence (AI) systems within the Judiciary of the Autonomous City of Buenos Aires. This resolution, adopted on February 26, 2025, reflects a proactive approach by the local government to integrate AI technologies into judicial processes while safeguarding fundamental rights and ensuring the integrity of the justice system. It acknowledges the transformative potential of AI to enhance efficiency in administrative and jurisdictional tasks, such as generating complex legal texts and assisting with case management, but also recognizes the inherent risks associated with its improper application, including the potential for errors or inaccurate results. The guidelines are designed to complement and augment human capabilities within the judiciary, emphasizing a human-centric approach where AI serves as a tool to support, rather than replace, human judgment and decision-making.

This regulatory instrument is a significant step in Argentina's broader effort to establish a robust framework for AI governance, building upon existing national data protection laws and drawing inspiration from international best practices. The resolution explicitly references various international and regional documents, including the Asilomar AI Principles, the EU's Ethics Guidelines for Trustworthy AI, UNESCO's Recommendation on the Ethics of AI, and the European Ethical Charter on the Use of AI in Judicial Systems. Furthermore, it considers other national and provincial initiatives, such as Argentina's Resolution 111/2024 on the National Integrated Artificial Intelligence Program in Justice and provincial protocols like San Juan's General Agreement No. 102/2024 (IAGen Protocol) and Río Negro's Accord No. 15/2024. The guidelines aim to provide a clear framework for judicial agents, promoting responsible innovation while mitigating risks related to privacy, discrimination, and the fairness of judicial outcomes. They underscore the importance of continuous training and education for judicial staff to ensure adequate understanding and ethical application of AI tools.

Definitions

Within the context of City of Buenos Aires Resolution 206/2025, several key terms are defined to ensure clarity and consistent application of the guidelines. "Artificial Intelligence Systems" (Sistemas de Inteligencia Artificial) generally refer to machine-based systems that, for given human-defined objectives, can make predictions, recommendations, or decisions influencing real or virtual environments. These systems are designed to operate with varying degrees of autonomy and can be applied across a wide range of judicial functions, from document analysis to predictive analytics. The guidelines specifically address "Generative AI Agents" (Agentes basados en inteligencia artificial generativa), which are a subset of AI systems capable of producing novel content, such as legal texts, summaries, or drafts, based on learned patterns from extensive datasets. The resolution emphasizes that the use of such generative AI must always be for professional and judicial purposes, strictly prohibiting personal use and the creation of offensive or illegal content.

The document also implicitly defines "Judicial Process" (Proceso Judicial) as the entire spectrum of activities undertaken by the judiciary, encompassing administrative tasks, legal research, case management, and ultimately, decision-making. The guidelines stress that AI systems are intended to complement and enhance these processes, not to replace the fundamental role of human judges and legal professionals. "Personal Data Protection" (Protección de datos personales) is defined in line with existing Argentine legislation, particularly Law 25,326, which remains the foundational statute governing AI-related data processing. The guidelines mandate strict adherence to these data protection norms, requiring anonymization of sensitive or confidential information before inputting it into AI systems and ensuring that AI applications respect privacy, confidentiality, and data integrity. Furthermore, concepts like "Bias" (Sesgo) and "Discrimination" (Discriminación) are understood as undesirable outcomes where AI systems might generate, reproduce, reinforce, or perpetuate unfair treatment, which the guidelines aim to prevent through careful design and oversight.

Governance and Institutional Framework

The governance framework for AI use in the Judiciary of the Autonomous City of Buenos Aires, as outlined in Resolution 206/2025, is primarily overseen by the Council of Magistracy (Consejo de la Magistratura). This body is responsible for issuing and enforcing the recommendations and guidelines, ensuring that the implementation of AI systems aligns with the principles of justice administration and the rule of law. The resolution underscores that the development, implementation, and use of AI models and generative AI agents must be human-centric, designed to complement and enhance the capabilities of the judicial branch while prioritizing respect for human dignity and autonomy. This institutional oversight aims to maintain the autonomy of judges and judicial officials in their decision-making processes, ensuring that AI tools serve as aids rather than substitutes for human judgment.

Beyond the Council of Magistracy, the framework implicitly calls for the establishment of internal protocols and designated personnel within judicial entities to manage AI adoption. This includes ensuring that judicial agents are adequately trained on the ethical, responsible, and secure use of AI technology. The guidelines also suggest a continuous analysis of proportionality when determining the type of AI model or agent to be implemented, considering the specific tasks, the level of impact on rights, and the degree of standardization of the tasks and decisions affected by the AI system. This implies a decentralized yet coordinated approach, where individual courts or departments may adopt specific AI tools, but their deployment must adhere to the overarching principles and oversight mechanisms established by the Resolution. The emphasis on human oversight (human-in-the-loop) is a cornerstone of this governance model, prohibiting the delegation of decision-making to AI without proper human review and validation of results.

Key Focus Areas

Resolution 206/2025 identifies several critical areas of focus to ensure the responsible integration of AI into the judiciary. A primary concern is Data Protection and Confidentiality, mandating that the use of AI systems strictly adhere to existing personal data protection regulations, particularly Law 25,326. The guidelines explicitly state that personal data or confidential information should, in principle, not be included in instructions given to AI tools, and users must anonymize data before entering it into AI systems to protect privacy and confidentiality. This focus extends to ensuring the integrity and security of information processed by AI, preventing unauthorized access or data manipulation that could compromise judicial processes or harm involved parties.

Another crucial area is Non-discrimination and Fairness, which seeks to prevent biased applications of AI systems and outcomes that could generate, reproduce, reinforce, perpetuate, or exacerbate discrimination. The guidelines require evaluating the implications of AI systems for procedural equity throughout their lifecycle, ensuring that implementations do not violate rights to justice and equal treatment. Transparency and Explainability are also paramount, requiring AI systems to be transparent, traceable, and auditable. Public entities using AI to interact with individuals must clearly inform users that such interaction is mediated by an AI tool, reinforcing principles of informational self-determination. This commitment to transparency extends to the verification of AI-generated results, with judicial agents explicitly prohibited from delegating decision-making to AI without proper human review. Finally, Risk Management is a core component, advocating for a preliminary risk assessment for all AI systems, categorizing risks based on their potential impact on safety and fundamental rights. The resolution introduces four risk categories—unacceptable, high, limited, and minimal/no risk—and prohibits systems deemed to pose an unacceptable risk, while high-risk technologies require algorithmic impact assessments and specific mitigation measures throughout their lifecycle. This proactive approach ensures that potential harms are identified and addressed before or during AI deployment.

Implementation Framework

The implementation framework for City of Buenos Aires Resolution 206/2025 is designed to facilitate the practical integration of AI systems into the judiciary while ensuring adherence to the established guidelines. A key aspect involves the development of internal protocols and operational procedures within each judicial entity to govern the specific use cases of AI. These protocols would detail how AI tools are to be selected, deployed, and monitored, ensuring consistency with the overarching principles of the resolution. The framework emphasizes a phased approach, where AI adoption is initially focused on tasks that complement human capabilities and enhance efficiency, such as data analysis, legal research, and document review, rather than automating critical decision-making functions. This measured implementation allows for continuous learning and adaptation, ensuring that the technology matures alongside the institutional capacity to manage it effectively.

Central to the implementation is a robust Education and Training program for all judicial staff. The resolution, in line with broader Argentine initiatives, recognizes the necessity of equipping judges, prosecutors, and judicial operators with the knowledge and skills to use AI ethically and effectively. Training initiatives would cover topics such as understanding AI capabilities and limitations, ethical considerations, data privacy protocols, and the critical importance of verifying AI-generated outputs. Furthermore, the framework encourages the creation of digital tools and platforms that support the ethical use of AI and enhance access to public information, fostering an environment of responsible innovation. The resolution also mandates ongoing identification and mitigation of bias or discriminatory outcomes, requiring continuous vigilance and adaptation of AI systems to prevent adverse impacts on fundamental rights. This proactive and adaptive implementation strategy aims to build trust in AI systems within the judiciary and ensure their long-term sustainability.

Monitoring and Evaluation

The monitoring and evaluation framework for Resolution 206/2025 is crucial for ensuring the ongoing effectiveness, ethical compliance, and adaptation of AI systems within the Judiciary of the Autonomous City of Buenos Aires. This framework necessitates continuous oversight of deployed AI tools to verify their adherence to the established guidelines, particularly concerning data protection, non-discrimination, transparency, and human oversight. Regular audits of AI systems are envisioned to assess their performance, identify any unintended biases or discriminatory outcomes, and ensure that they continue to operate within the defined ethical and legal parameters. These audits would likely involve both internal reviews conducted by designated judicial bodies and potentially external evaluations to provide an independent assessment of compliance and impact.

Furthermore, the evaluation process extends to assessing the impact of AI on fundamental rights and the overall integrity of the judicial process. This includes analyzing how AI tools affect procedural fairness, access to justice, and the autonomy of human decision-makers. The resolution's emphasis on a proportionality analysis for AI deployment suggests an ongoing need to re-evaluate whether the benefits of an AI system outweigh its potential risks in specific contexts. Feedback mechanisms from judicial staff, legal professionals, and even the public are likely to be incorporated to gather insights on the practical application and perceived effectiveness of AI tools. This iterative monitoring and evaluation process allows the Council of Magistracy to adapt the guidelines as technology evolves, new risks emerge, or best practices are refined, ensuring that the regulatory framework remains relevant and robust in the face of rapid AI advancements. The Provincial Registry of Artificial Intelligence Systems, established by a similar resolution in the Province of Buenos Aires, provides a precedent for tracking and overseeing AI systems in public entities, suggesting a similar mechanism could be adopted for the City's judiciary.

Penalties, Liability, and Appeals

While City of Buenos Aires Resolution 206/2025 primarily provides guidelines and recommendations, the implications for non-compliance with its principles are significant and would likely draw upon existing legal frameworks within Argentina. The resolution's emphasis on responsible and ethical AI use, particularly regarding data protection, non-discrimination, and human oversight, implies that violations could lead to various forms of accountability. For instance, the misuse of personal data by AI systems, or the failure to anonymize sensitive information, would fall under the purview of Argentina's Personal Data Protection Law (Law 25,326), which carries its own set of penalties for non-compliance. Judicial agents who manipulate data or use AI to spread content that violates laws or ethical standards, as highlighted in other Argentine protocols, could face disciplinary actions within the judiciary, ranging from formal admonishments to more severe sanctions.

In terms of liability, the guidelines reinforce the principle that human oversight remains mandatory and that AI cannot replace human decision-making, particularly in sensitive areas like justice. This suggests that ultimate responsibility for judicial decisions, even those informed by AI, rests with the human judge or legal professional. Therefore, any errors or harms resulting from AI use, especially if due to a lack of proper verification or oversight, would likely lead to accountability for the human agent involved. The resolution implicitly establishes a high standard of due diligence for judicial personnel utilizing AI tools. Regarding appeals, decisions made with the assistance of AI would remain subject to the existing appellate processes within the Argentine judicial system, allowing for review and challenge based on legal and factual grounds. The ability to appeal decisions influenced by AI is a crucial safeguard for fundamental rights, ensuring that individuals can seek redress if they believe an AI system has contributed to an unfair or incorrect outcome.

Relationship to Other Instruments

City of Buenos Aires Resolution 206/2025 does not exist in a vacuum but is intricately linked to a broader ecosystem of legal and policy instruments at national, provincial, and international levels. At the national level, the resolution is founded upon and complementary to Argentina's foundational Personal Data Protection Law No. 25,326, which remains the core statute governing any AI activities involving personal data. This ensures that AI systems deployed in the judiciary respect existing privacy rights and data handling obligations. Furthermore, it aligns with national initiatives such as the National Integrated Artificial Intelligence Program in Justice (Resolution 111/2024), which aims to accelerate administrative and judicial procedures while explicitly protecting fundamental rights and expanding access to justice. The resolution also considers the ethical principles for AI human oversight, accountability, and transparency set out in national guidelines like Provision 2/2023 and the Guide for Public and Private Entities on Transparency and Protection of Personal Data for Responsible AI (2024).

On a provincial level, the Buenos Aires Resolution draws lessons and inspiration from other regional protocols adopted by judicial authorities in provinces such as San Juan (General Agreement No. 102/2024, the IAGen Protocol) and Río Negro (Acordada N° 15/2024). These provincial protocols provide concrete examples of how generative AI is being regulated, often mandating strict data anonymization, prohibiting inappropriate content, and requiring human verification of AI outputs. The resolution also fits within the context of the City of Buenos Aires' own "Strategic Artificial Intelligence Plan - Future City," developed in 2021, which promotes an integral and multi-stakeholder approach to AI development. Globally, the guidelines explicitly consider and integrate principles from international instruments such as the Asilomar AI Principles, the EU's Ethics Guidelines for Trustworthy AI, the UNESCO Recommendation on the Ethics of AI, and the European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems. This demonstrates a concerted effort to align local regulations with global best practices and emerging international standards for AI governance, including the EU AI Act (Regulation 2024/1689), which is explicitly referenced.

International Alignment

City of Buenos Aires Resolution 206/2025 demonstrates a strong commitment to international alignment in its approach to AI governance within the judiciary. The resolution's preamble and considerations explicitly reference a wide array of international principles, guidelines, and regulatory frameworks that informed its drafting. This includes foundational documents such as the Asilomar AI Principles, which advocate for beneficial AI, and the European Commission's Ethics Guidelines for Trustworthy AI, emphasizing human agency, technical robustness, privacy, transparency, and accountability. By incorporating these widely recognized ethical frameworks, the City of Buenos Aires positions its guidelines in harmony with global efforts to ensure AI development and deployment are human-centric and value-driven.

Furthermore, the resolution specifically cites the UNESCO Recommendation on the Ethics of Artificial Intelligence and the UNESCO Global Toolkit on AI and the Rule of Law, underscoring an adherence to principles of human rights, non-discrimination, and sustainable development in the context of AI. The European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems, developed by the Council of Europe's European Commission for the Efficiency of Justice (CEPEJ), is also a key reference, highlighting the focus on judicial-specific ethical considerations such as impartiality, fairness, and the right to a fair trial. The mention of the EU AI Act (Regulation 2024/1689) indicates an awareness of leading global regulatory models, particularly its risk-based approach and emphasis on transparency and human oversight. This comprehensive referencing of international instruments ensures that the Buenos Aires guidelines are not only locally relevant but also globally compatible, facilitating potential future international cooperation and mutual recognition in the evolving landscape of AI regulation.

Implementation Timeline

Sources and References