Compliance

Auditability of the AI system

Capacity to evaluate algorithms, data, and design or undergo audits.

What it triggers

  • Under Decree 98/2025 of the Principality of Asturias, deployers are required to provide access to documentation, logs, and other evidence necessary for independent review and verification of AI systems.source

Definition

The capacity to evaluate an AI system's algorithms, data, and design processes, including the ability to undergo formal auditing procedures to assess compliance, performance, and impacts. It encompasses access to documentation, logs, and other evidence necessary for independent review and verification.

Not to be confused with

Audit Trail

Audit trail refers specifically to immutable logs of inputs, prompts, outputs, and system versions, while auditability encompasses the broader capacity to evaluate algorithms, data, design processes, and undergo formal audits.

AI Auditing

AI auditing describes the processes for examining AI systems for ethics, accuracy, and compliance, whereas auditability is the inherent capacity of the system to be subjected to such audits.

Questions and answers

What does auditability require for AI systems in Asturias public sector?

Auditability refers to the capacity to evaluate an AI system's algorithms, data, and design processes, including the ability to undergo formal auditing procedures, with access to documentation, logs, and other evidence for review.

How does auditability differ from an audit trail?

Auditability is the overall capacity of an AI system to be evaluated and audited, including access to necessary documentation and logs. An audit trail specifically provides immutable logs of inputs, prompts, outputs, and system versions as one component supporting that capacity.