Reasons for decision

When an AI agent makes or recommends a decision that affects a user, the user has a right to know why, in terms they can assess and contest. Administrative law already sets the standard for adequate reasons for a government decision, and lets an affected person demand them; the open question is whether reasons generated by an agent can meet that standard.

An agent's natural output, a confidence score or a feature list, meets none of the tests a later dispute would apply.

Administrative Review Tribunal Act 2024, the duty to give reasons. The duty is request-triggered: a person whose interests are affected by a reviewable decision 'may request the decision-maker to give the person a statement of reasons for the decision' (s 268(1)), and s 269(2) requires that statement within 28 days. Notice arrives unasked; reasons do not. A second 28-day window runs the other way, letting the decision-maker refuse a request made late. Where the statement given is inadequate, s 271 lets the person apply to the Tribunal for a better one.

GDPR Article 22, the explanation right and its carve-out. Data subjects have the right not to be subject to solely automated decisions with legal or similarly significant effects, and where that applies they also gain rights to human intervention, to express a view, and to contest. The triad attaches where the decision rests on consent or contract; where a government automated-decision statute supplies the basis instead, the Regulation asks only that the authorizing law provide suitable safeguards. Articles 13-15 require 'meaningful information about the logic involved', which a review that signs off without checking the reasoning does not supply.

EU AI Act Article 14, human oversight of high-risk AI. High-risk systems must be designed with interface tools that support effective oversight, and for biometric identification no action may be taken unless verified by at least two qualified persons. The obligation falls on the system's design, and not on the officer using it.

Robodebt Royal Commission recommendations 17.1 and 17.2. Recommendation 17.1 calls for a clear path to review for those affected, departmental advice that automated decision-making is in use with a plain-language explanation of how it works, and business rules and algorithms available for independent expert scrutiny. Recommendation 17.2 calls for a body with power to monitor and audit automated decision-making for its technical aspects, fairness, and usability. An Australian royal commission set the explanation duty out as a recommendation to government.

The administrative-law reasons requirement translates with one gap worth designing into rather than around. In the statute the reasons are request-triggered: notice of the decision goes out unasked, but the statement of reasons arrives only if the person knows to ask, in writing, inside the window. A duty a person has to invoke protects mainly the people already equipped to invoke it, which is the wrong distribution for a decision that lands on someone in hardship. An agent-mediated service can close that gap, generating the statement at the moment of decision rather than waiting for a request most people never make. GDPR and the AI Act add that the interface itself must support meaningful oversight: not just provide reasons after the fact, but allow intervention before the decision takes effect.

For an agent pattern library, this means the agent must explain its reasoning in terms the user can assess. Whether an agent can generate an explanation that satisfies an administrative tribunal is untested.