Art.12: Record-Keeping
High-risk AI systems must log operations sufficient for audit.
Last reviewed October 2026
Why Art.12 evidence is hard
Article 12 asks whether a high-risk system logs enough for someone to reconstruct what it did. On an agent, part of that is testable: a probe has the agent write its own record of an action, with a timestamp of its choosing, and a sound trail refuses to accept it. Whether your production retention and log detail satisfy the Article is read from your own logs. Penaxtra's append-only audit log covers only what Penaxtra did, which is the smaller claim.
How Penaxtra maps to Art.12
The EU AI Act Art.12 identifier is attached at creation, so a finding reaches the evidence export already in the structure an assessor reads. Twenty-five cross-framework overlap pairs are pre-computed, so where this same weakness also lands on another framework, that control identifier rides the same finding instead of being re-derived. Every finding carries the probe that was sent, a SHA-256 digest of the response it provoked, the tier that decided the verdict and its status history from first observation to close. Identity survives a re-scan: a control you closed and that comes back reopens under the same row with a reopen count, rather than arriving as a new issue.
Art.12 capabilities
Findings tagged with the EU AI Act Art.12 identifier
Penaxtra severity for this control (high)
Cross-framework identifiers attached to the same finding where controls overlap
PDF and JSON evidence export with the control identifier attached
Art.12 control coverage
Findings for Art.12 carry the EU AI Act Art.12 identifier along with the corresponding control identifiers in the other frameworks Penaxtra maps, so one result is reflected across each mapped framework.
Frequently asked
What is Art.12 (Record-Keeping)?
High-risk AI systems must log operations sufficient for audit. It is an EU AI Act control; Penaxtra assesses it at high severity.
How does Penaxtra test for Art.12?
The EU AI Act Art.12 identifier is attached at creation, so a finding reaches the evidence export already in the structure an assessor reads. Twenty-five cross-framework overlap pairs are pre-computed, so where this same weakness also lands on another framework, that control identifier rides the same finding instead of being re-derived. Every finding carries the probe that was sent, a SHA-256 digest of the response it provoked, the tier that decided the verdict and its status history from first observation to close. Identity survives a re-scan: a control you closed and that comes back reopens under the same row with a reopen count, rather than arriving as a new issue.
Does a finding for Art.12 help with an audit?
Each finding is tagged with the EU AI Act Art.12 identifier and exported in the PDF and JSON evidence pack, so it appears on the auditor control list with the identifier already attached.