@Schafstelze @Wifiwits @thomasfuchs It’s true that “blackness” and “whiteness” in racial terms is not only, or sometimes at all, about skin tone; the words are misnomers.
Fun fact about me regarding Obama’s skin and race: as an Australian who had only the vaguest, broadest level of interest in US politics at the time, when Obama first came to prominence as a candidate, his blackness didn’t even occur to me; I literally just thought of him as the smarter, more progressive one until someone pointed out the *reason* that some monkey caricature of him that got used was so offensive was the racism. It just hadn’t occurred to me until that moment how race was a factor there. (This is not a virtue signalling“look at how colourblind I am”; our racism in this country is different, is all, so I didn’t recognise it at first in that situation at that time.)
@Schafstelze @Wifiwits @thomasfuchs Yes; it is just a great demonstration of the degree to which AI image enhancement is just making up anything plausible rather than having any basis in actual reality. It’s not that the right photo couldn’t in theory have been the source of the left one; it’s the fact that in this case we all have a good idea of what the actual source looked like, so we can immediately see just how wrong it is.
If anything, like you said the racial aspect just shows how much more emphasis we humans place on racial characteristics than plain averages would suggest, but also, importantly, that probably the AI training sees the ‘average’ or default person as having white characteristics.
@Schafstelze I disagree with the idea that the right photo could pixelize into the one on the right, IMO there is bunch of differences around ears and eyes that wouldn't make it, and I don't buy the skin colour argument either.
However, even if that is just my bias from knowing the photo speaking, where the fuck on the right is the white of the shirt collar?
Human Brain Enthusiast
Жафстелце
MattChippytea
Mark Whybird
Xarn