In the sphere of music we observe that autotune impacted the voice of new generations of singers, making them align more towards the fake/uncanny valley, instead of autotune becoming better to get out of the uncanny valley. This effect is hypothesized to come from the abundance of autotuned voice exposure new singers impress and train their ears with starting from early age.

This generates a self feeding loop which on a societal level shifts the location of the uncanny valley. Eventually, to the vast majority of people, old not-autotuned music will sound uncanny or off.

Adam Neely made a interesting episode about this: https://youtu.be/yxX2u8iggYI (around 5:00)

I wonder if we will observe something similar in acting and social interaction, where more and more bad or CGI acting gets corrected by uncanny valley methods, which new generations get on mass exposed with and thus change their own habits and what it means for facial expressions to be uncanny.

There is an argument to be made that due to economic reasons, eventually it will not pay off to invest into reaching 100% out of uncanny valley but rather a “good enough” implementation will gain mass traction and thus change perception of what the uncanny valley is. Old movies will be the uncanny valley.