Transparency and explainability

AI-related objective: Appropriate information about the AI system and the reasons for its outputs is available to the stakeholders who need it, at the level and in the form they need. Used when establishing risk criteria (6.3.4), identifying consequences (6.4.2.6) and evaluating risk (6.4.4). Read transparency and explainability against the objective the use threatens most.

What it asks

AI-related objective: Appropriate information about the AI system and the reasons for its outputs is available to the stakeholders who need it, at the level and in the form they need. Used when establishing risk criteria (6.3.4), identifying consequences (6.4.2.6) and evaluating risk (6.4.4).

Where generative AI changes the answer

Transparency means a person can learn that a generative system was used, what it was asked and what limits apply, in language they can act on.

Reading it in practice

Transparency for a generative AI system is not a model card alone. The evidence is a plain statement, in the user's language, that the system is generative, what it was asked to do, what it cannot do and how a person can question an output. Test it by asking a user to explain the limits after reading it. If they cannot, the statement is decoration.

Who it binds

The AI risk owner, with the objective owner named in the risk treatment plan. For A.12 the named role is the owner of documentation of the system and explanation capability suited to each stakeholder.

Evidence an auditor normally samples

• Documentation of the system and explanation capability suited to each stakeholder

The test you can run

Ask for documentation of the system and explanation capability suited to each stakeholder and for its last review date; an artefact nobody can date is paperwork rather than a control.

Common failures

• No explanation available where decisions affect people

Before the draft publishes

For A.12, pick the two objectives the use most affects, write the test you would apply, and record the result before the draft is published. The first artefact to produce is documentation of the system and explanation capability suited to each stakeholder.

Where this sits against the draft

A.12 carries the annex a: ai-related objectives part of the draft's published scope, which names risk sources, the stakeholders exposed to them, risk analysis, risk treatment and controls. When the draft publishes, compare its wording on A.12 (transparency and explainability) with this record and note the change in the register.