Feature summary (what you would like to be able to do and where):

As a follow-up to T435136, implement a feature for image captions that are drawn from the Wikipedia article, the Commons description, or are user-generated. Especially significant when we are illustrating the concept by something related that isn't the subject itself (e.g. a photo of a stadium from 5 years ago for a current sports tournament.)

Use case(s) and Benefits (Describe the actual underlying problem which you want to solve. Why should this be implemented? Do not describe only a solution):

Provides more content to the image, and makes it less misleading if it's a related image, rather than a direct depiction of the news story.