The minor hue change of the picture makes them both published for "substantially the same purpose" i.e. subject-aggrandizement for being in position X at time Y

The minor hue change of the picture makes them both published for "substantially the same purpose" i.e. subject-aggrandizement for being in position X at time Y. AI training doesn't preserve position nor the purpose of a photograph's subject like photoshop does, it amasses a style from corpus as probabilities. Upstream it's not substantial copying of a single work, downstream it's substantial restyling into any new imagined purpose. Slam dunk.