automated decision‑making

Decisions or assessments in which algorithmic outputs materially inform or determine outcomes for a citizen, such that the algorithm's result plays a substantive role in the administrative decision.

A class of AI-driven processes in which decisions are executed by a system without real‑time human intervention or review; the report contrasts this with decision‑support tools that assist human decision‑makers and may require different supervisory expectations and potential restrictions for high‑risk, fully automated uses.

The use of algorithmic or AI systems to reach decisions that materially affect individuals without meaningful human intervention; where such decisions affect rights or significant interests, controllers must provide specific notice and safeguards such as human review and contestability.

A clarified definition distinguishing fully automated from assisted decision processes, setting out when decisions qualify as ADM under the law and the associated obligations (notice, representation, human review pathways and safeguards) particularly for non‑sensitive data where wider automated processing is permitted.