Dont know about that incident, but that is different than scanning for csam which is different than scanning for faces which is different from feeding your images into an genAI training set.
You could think that they are doing all this anyway (which I think the AI only companies do btw, vut doubt the bigger ones do, esp Apple).
This has been a problem for a while esp with podcast providers, who put ice ads in podcasts and then the podcasts get blamed. (Despite having very little control over it, and the control also often being ignored).