LDR.Reference:5 - People, organizations and capable AI

Human learning and agent engineering require different methods. HCD applies when the subject is a human capability. An organization supplies a contribution through its arranged people, agents, resources and authority; it does not acquire a human skill simply because its members attended training. An AI system’s qualification, configuration, access and operational support need their own evidence and engineering.

A capable AI may produce the whole leadership contribution, including substantive integration or a decision delegated to it. Do not freeze the old arrangement and merely insert a drafting assistant. Compare whole alternatives: ordinary human coordination, a person with AI support, autonomous agents under effective mandates, or a changed allocation that removes a previously necessary meeting or intermediate role. Establish actual competence for the task, permitted information use, ability to make the relied-on commitments, operating capacity, exception handling and a receiver able to use the result.

Cheap computation does not make attention, transition, authority, verification or another participant’s work free. Nor does it justify preserving human duplication when that work provides no needed guarantee. The constructed application below shows how the comparison can select autonomous performance under one set of conditions and restore a narrower human contribution after a consequential change.