I worked on Google Image Search, and my team was responsible for image quality across all of Search. There are some checks on image similarity, although “de-dup

I worked on Google Image Search, and my team was responsible for image quality across all of Search. There are some checks on image similarity, although “de-duplication” is applied in smart ways for different queries. To my knowledge, there isn’t an algorithm that tries to pick “non-stock” images. In general, Google/page rank favors sites that other people have favored (by linking to and talking about on the internet). It seems that original content is generally more “link-worthy” than copied content, so that might explain your observation.