Technical

Audit Trail

Immutable logs of inputs, prompts, outputs and system versions.

How the laws define this

Both instruments agree that an audit trail is a tamper-evident record essential for enabling audits and accountability. Where they diverge is in the scope of the system activities that must be logged. Under the Resolución de Presidencia 206/2025 – Guía de Recomendaciones y Directrices para el Uso de Sistemas de Inteligencia Artificial en el Poder Judicial de la Ciudad Autónoma de Buenos Aires, audit trails focus on operational interactions, capturing inputs, prompts, system responses, model versions, and relevant metadata to enable the reconstruction, review, challenge, and independent auditing of AI-assisted processes. In contrast, the AI and Data Protection Regulation Bill mandates comprehensive coverage across the entire AI system lifecycle, requiring records of development and governance events—such as dataset versions, training runs, model updates, testing results, and deployment decisions—alongside decision outputs to support regulatory compliance investigations.

Synthesised from the 2 statutory definitions below. Each one is quoted in full, with its source.

How jurisdictions differ

What is distinctive in each definition, binding law in force first.

Jurisdiction	Instrument	Status	What is distinctive
Argentina	Presidential Resolution 206/2025 – Guide of Recommendations and Directives for the Use of Artificial Intelligence Systems in the Judiciary of the Autonomous City of Buenos Aires	In Force	This definition emphasizes judicial processes by guiding the maintenance of audit trails to support review, challenge, and independent audits of operationally used AI systems.
Ecuador	AI and Data Protection Regulation Bill	Under Review	This definition widens the scope to the full AI system lifecycle, including datasets, training runs, model updates, testing results, and deployment decisions.

What it triggers

In the Judiciary of the Autonomous City of Buenos Aires, the Guide recommends maintenance of audit trails to support review, challenge, and independent audits of systems used operationally.source
Under Ecuador's AI and Data Protection Regulation Bill, if enacted it would require comprehensive audit trails to enable audits, accountability, and regulatory compliance investigations.source

Definitions (2)

A secure, tamper-evident record containing inputs, prompts, system responses, model versions and relevant metadata to enable reconstruction, auditing and accountability of AI-assisted processes. The Guide requires maintenance of audit trails to support review, challenge, and independent audits of systems used operationally.

Resolución de Presidencia 206/2025 – Guía de Recomendaciones y Directrices para el Uso de Sistemas de Inteligencia Artificial en el Poder Judicial de la Ciudad Autónoma de Buenos AiresDefinition 1 of 2

Comprehensive, tamper-evident records capturing key events across an AI system's lifecycle — including dataset versions, training runs, model updates, testing results, deployment decisions, and decision outputs — to enable audits, accountability, and regulatory compliance investigations.

AI and Data Protection Regulation BillDefinition 2 of 2

Not to be confused with

data use audit trail

Data use audit trail is limited to automated logging of dataset access and processing activities, while audit trail covers broader inputs, outputs, and system events.

Record of processing and design decisions

Record of processing and design decisions focuses on documented lifecycle logs for auditability, whereas audit trail includes tamper-evident records of inputs, prompts, and outputs.

Questions and answers

What does Argentina's guideline require for audit trails in judicial AI use?

The Guide of Recommendations and Directives for the Use of Artificial Intelligence Systems in the Judiciary of the Autonomous City of Buenos Aires recommends maintenance of audit trails containing inputs, prompts, system responses, model versions and metadata. This supports review, challenge, and independent audits of AI-assisted processes.

What lifecycle events must Ecuador's bill cover in audit trails?

The AI and Data Protection Regulation Bill proposes comprehensive records capturing dataset versions, training runs, model updates, testing results, deployment decisions, and decision outputs. These enable audits, accountability, and compliance investigations.

How do audit trails differ between judicial guidelines and proposed AI bills?

Argentina's guideline focuses audit trails on operational judicial processes for review and audits. Ecuador's bill expands to full lifecycle events for broader regulatory compliance.