I’m not familiar with Google here, but Apple is quite loquacious in its descriptions of how it approaches computational photography, with extensive use of bracketing and sensor fusion – and nothing at all about generative embellishments.
Not even saying it’s impossible, but I would be absolutely shocked if Apple were actually using anything outside a lot of bracketed exposures and optically-modeled diffs between multiple sensors. More than anything, it would be distinctly off-brand.
They put very obvious fake bokeh in their photos, and their users seem happy about it. I don't know what you mean by "generative" and how you'd draw a line between "generative" and not, but they're absolutely not giving you the straight-up picture you took.
I don't they're doing a fake milky way, but I don't think that they'd be above it if they could come up with a reason to.
https://www.fastcompany.com/90372724/welcome-to-post-reality...
"X did Y"
And when it turns out they didn't:
"It doesn't matter because they WOULD"
(Disclaimer: I have no idea if Apple does or doesn't do the milky way thing)
I'm saying X did Y, which they did, and it means that Z is at least on the table, where X = Apple, Y = FaceTime Eye Contact feature, Z = Milky Way enhancement.
Where have you seen this? I can't find anything about it online, and I imagine it would be even more egregious than the Samsung fake moon, probably putting it in the wrong place or orientation in the sky sometimes.
I'm sure the iPhone is just as capable
It recognized the moon and replaced the real moon taken by the camera with a stored stock photo pasted over it.
There was no “upscaling” involved.
It's like when Ryan Gosling faces started being hallucinated into generated images because of overfitting of his image in the training data set