I suspect that there are people on the panel who know more about the technology than I do, but if algorithms are trained on a huge set of images, and one of those images, or more than one, is of you, then the cake is already baked. It's difficult to unbake the cake and remove the influence of any one individual from the database on the algorithm that emerges from it.
On April 4th, 2022. See this statement in context.