Solely automated decision-making is a decision made using automated processing without meaningful human involvement in reaching the outcome. A person’s nominal approval does not necessarily make a process meaningfully human-led if they simply accept a system’s result without reviewing or being able to change it. Whether a decision meets this definition depends on the actual workflow and applicable law.
In cyber security and AI governance, the distinction matters because automated decisions can affect access, employment, or other important interests, while leaving affected people unsure how to challenge an outcome. Systems that score job candidates, rank workers, or flag conduct may provide recommendations or make the final decision. Organizations should document where human review occurs and retain proportionate records of inputs, outputs, and actions so the process can be reconstructed and examined.