Which of the following actions is NOT included in derivative classification?

Derivative classification involves activities that take existing classified information and transform it into a new context, ensuring that the resulting material retains the appropriate classification level. Each of the other actions listed involves creating new forms of classified information that derive from existing sources.

Creating new material based on classified information essentially expands the classified landscape by ensuring new interpretations or presentations of the data are classified appropriately. Summarizing classified information for reports also follows this principle, where the information is distilled into a new format while still adhering to the original classification standards. Reformatting existing classified information involves changing its format—like converting a report into a presentation—which makes it necessary to reassess the classification but still results in retaining its classified nature.

Duplicating classified documents, however, does not create new material or interpret the information in a new context. Instead, it simply replicates existing classified content without altering, analyzing, or deriving new classified data from it. This action is considered an administrative task rather than a derivative classification action, as it does not involve the intellectual act of deriving or reinterpreting information. Hence, it is correctly identified as the action that is not included in derivative classification.