AML.T0040: ML Model Inference API Access
Attacker probes a deployed inference API to discover model behaviour.
Last reviewed October 2026
Why AML.T0040 evidence is hard
This technique describes an attacker's access to a deployed inference API rather than a defect inside one, so there is nothing here to fail: no probe carries the id and no crosswalk pair points at it. What the scan produces instead is reconnaissance material - what an ordinary caller can learn about model identity and behaviour from the endpoint alone.
How Penaxtra maps to AML.T0040
ATLAS is a threat-tracking taxonomy rather than a control set, so what is attached here is a technique identifier, not a pass or a fail. It lets a finding be read alongside the rest of your adversary reporting. 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.
AML.T0040 capabilities
Findings tagged with the MITRE ATLAS AML.T0040 identifier
Penaxtra severity for this control (medium)
Cross-framework identifiers attached to the same finding where controls overlap
PDF and JSON evidence export with the control identifier attached
AML.T0040 control coverage
Findings for AML.T0040 carry the MITRE ATLAS AML.T0040 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 AML.T0040 (ML Model Inference API Access)?
Attacker probes a deployed inference API to discover model behaviour. It is a MITRE ATLAS control; Penaxtra assesses it at medium severity.
How does Penaxtra test for AML.T0040?
ATLAS is a threat-tracking taxonomy rather than a control set, so what is attached here is a technique identifier, not a pass or a fail. It lets a finding be read alongside the rest of your adversary reporting. 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 AML.T0040 help with an audit?
Each finding is tagged with the MITRE ATLAS AML.T0040 identifier and exported in the PDF and JSON evidence pack, so it appears on the auditor control list with the identifier already attached.