Now I agree that the model photo looks a little more artificial for reasons beyond gender, and they should probably have compared with a male model photo as well. Though personally I suspect people would still think it more likely that an attractive man was a real person than an attractive woman.
Except it's not. Stock photos of good looking women are routinely used in marketing to attract users. It's proven by stats from A/B testing.
Okay, I'm being glib, but this idea that thinking of a plausible causation with no further critical analysis means it's "not sexism" is kind of exasperating.
I'm only saying a person using good looking woman as an avatar because of a/b testing does not make him sexist.
Whether they make a decision based on internal biases caused by societal norms, or make a decision by measuring those societal norms directly and responding to them, the decision is still an expression of societal sexism.
They're photos, though. They're real. Not "fake"