Just a couple of years ago, when AI generation of images and videos became publicly available, we talked about the uncanny valley effect as a difficult-to-bridge gap between an artificial image and a real video/photo. However, today, advertising experts who actively use generative networks to create advertisements talk about an increase in the attractiveness of artificial images and an increase in sales of those products in the advertisement of which the presence of a neural image is obvious.
How can this be interpreted? AI-images still differ from real ones, people-images in them have not become indistinguishable from real ones. Can we say that we have simply gotten used to the unnaturalness of generated portraits, have stopped being afraid of them, as we once got used to the moving train from the Lumiere brothers' film? It turns out that the "uncanny valley" is not a death sentence? We just had to get used to it? Are you aware of the latest research on this topic that would say that the uncanny valley problem is far-fetched?